AI Innovations in Risk Management
A Case Study of Volvo AB’s Supply Chain Resilience
Authors:
Emma Hiljemark
Donika Nika
Supervisor: Kevin Cullinane
Master’s Degree Project in Logistics and Transport Management
Spring 2024
Graduate School, School of Business, Economics and Law, University of Gothenburg
AI Innovations in Risk Management - A Case Study of Volvo AB’s Supply Chain Resilience
AI Innovations in Risk Management - A Case Study of Volvo AB’s Supply Chain Resilience
By Emma Hiljemark & Donika Nika
© Emma Hiljemark Donika Nika
School of Business, Economics, and Law, University of Gothenburg
Vasagatan 1, P.O. Box 610, SE 405 30 Gothenburg, Sweden
Institute of Industrial and Financial Management & Logistics
All rights reserved.
No part of this thesis may be distributed or reproduced without written permission
by the authors.
Contact: emma.hiljemark@gmail.com ; donikanika@hotmail.com
2
Acknowledgement
The significance of this research lies in its contribution of valuable insights aimed at
enhancing strategic decision-making processes and fostering the development of more
resilience and robust supply chain systems for Volvo AB. Moreover, it seeks to advance the
understanding of how artificial intelligence can be effectively leveraged to enhance supply
chain risk management.
We would like to extend our sincere gratitude to Kevin Cullinane, whose exceptional
supervision has been invaluable throughout this journey. Your stimulating discussions,
availability, and innovative insights have greatly enriched this master thesis.
Furthermore, we express our heartfelt appreciation to Volvo AB for their collaboration. The
discussions, meetings and interview conducted with them have formed the thesis, providing it
with depth and substance. Your support throughout this research has been invaluable and
greatly appreciated.
3
Abstract
This research aims to explore the potential of various artificial intelligence technologies in
enhancing risk management within Volvo AB’s supply chain management. Application areas
include forecasting, procurement, inventory planning and management, and the development
of supply chain networks. Through an examination of the contributions of AI in these
domains, this study seeks to give an overview of valuable insights into decision-making
processes, operation and cost efficiency, mitigation of risks in the supply chain and fostering
competitive advantages for companies. The study is based on a qualitative approach with a
comprehensive literature review that provides an overview of advanced industry 4.0
technologies such as blockchain, internet of things, explainable artificial intelligence and
cloud computing. These technologies hold considerable promise within different areas in the
supply chain, due to their capacity to analyze extensive datasets effectively and discern
patterns, thereby enabling more precise and accurate predictions of potential disruptions and
risks. Furthermore, a case study with interviews involving expertise in the area from Volvo
AB has been conducted to analyze the challenges and opportunities of AI adoption in risk
management practices at Volvo AB.
The concluding remarks of this research includes that Volvo AB must recognize both
opportunities and implications in AI implementation. Integration of advanced technologies
like artificial neural network (ANN), cloud computing and blockchain enhances resilience,
efficiency and transparency within the supply chain. Specifically ANN, presents itself as a
promising tool in identifying patterns, forecasting results and streamlining operations for
Volvo AB. A successful implementation requires a comprehensive strategy including
regulatory frameworks, change management and resource allocation. A strategic and holistic
approach to AI implementation will aid Volvo AB navigate risks and maximize advantages,
solidifying its leadership in the market.
Keywords: supply chain, disruptions, industry 4.0, supply chain risks, AI, risk management,
supply chain risks, risk mitigation.
4
Table of Content
Acknowledgement 3
Abstract 4
1. Introduction 7
1.1. Background 7
1.2. Problem Statement 8
1.3. Research Questions 8
1.4. Significance of the Study 8
1.5. Delimitations 9
1.6. Theoretical Framework 9
1.6.1. Risk Management 9
1.6.2. Disruptions in the Supply Chain 10
1.6.3. Enterprise Risk Management System 13
2. Literature Review 14
2.1. Application Areas of AI in Supply Chain Management 14
2.1.1. Forecasting 14
2.1.2. Procurement 15
2.1.3. Inventory Planning and Management 18
2.1.4. Supply Chain Network 18
2.2. Industry 4.0 Technologies 20
2.2.1. Artificial Intelligence 20
2.2.1.1. Advantages of Artificial Intelligence 21
2.1.1.2. Challenges with Artificial Intelligence 21
2.2.1.3. Artificial Intelligence of Things (AIoT) 23
2.2.1.4. Explainable Artificial Intelligence (XAI) 24
2.2.2. Machine Learning 26
2.2.3. Cloud Computing 26
2.2.4. Internet of Things (IoT) 27
2.2.5. Artificial Neural Networks (ANN) 28
2.2.6. Agent-Based System 30
2.2.7. Blockchain 31
2.2.7.1 Blockchain in Risk Management 32
2.2.8. Summary of Industry 4.0 Technologies 33
2.3. Supply Chain Risks 39
2.3.1. Regulatory and Compliance Risks 39
2.3.2 Supplier Risks 40
2.3.2.1. Monitoring of Suppliers 43
2.3.2.2. Prediction and Identification of Supplier Risks 44
2.3.3. Resilience Planning and Strategy 45
2.4. Implementation Strategies for AI 46
2.5. Gaps in the literature 48
5
3. Methodology 49
3.1. Research Design 49
3.2. Data Collection 50
3.2.1. Validity and Reliability 51
4. Empirical Findings 51
5. Discussion 53
6. Analysis 57
6.1. AI technologies in Volvo AB’s Supply Chain 57
6.2. AI in different application areas for Volvo AB’s supply chain 58
6.3. AI implementation Risks and Strategy for Volvo AB 59
6.4. AI possibilities for Volvo AB in the future 61
7. Conclusions 62
7.1. Limitations & Future Research 63
References 65
Appendices 78
Appendix A - Interview Questions 78
6
1. Introduction
This chapter lays the fundamentals for this thesis, covering essential elements such as
background, problem statement, research questions, significance, delimitations, and the
theoretical framework, which prepares the reader for the upcoming chapters.
1.1. Background
Due to globalization, supply chains have become more exposed and vulnerable towards
disruptions and risks (Duong Khanh & Chong, 2020). Disruptive risks often stem from areas
of supply, production and transportation in the supply chain (Ivanov et al., 2017) and other
identified sources of disruptions encompass incorrect forecasts, natural disasters and delays
(Ivanov, 2021). The impact of disruptions can cause severe consequences for the organization
involving decreases in revenues, lowered product quality and increased supply chain and
shortage costs (Kassa et al., 2023). Artificial intelligence (AI) holds the ability to support
dynamic capabilities of risk mitigation within the supply chain and allows the organization to
increase their operational efficiency and accuracy within predictions of demand (Belhadi et
al., 2021; Aliahmadi et al., 2022). It is able to provide optimized solutions for issues
regarding for instance new regulations and global challenges through reduced costs, speed,
agility, efficiency and flexibility (Adeyemi et al., 2023). Implications of implementing
artificial intelligence within the supply chain include relying heavily on the software and
technical infrastructure, thus risks of providing unreliable results if programmed incorrectly.
Furthermore, AI is most suitable for narrow supply chain problems and may not be able to
handle uncertainties and risks regarding more complex decisions (Min, 2010).
Volvo AB was founded in 1927, and is based in Gothenburg in Sweden. Volvo engages in
more than 190 market operations where the majority of Volvo AB's customers are companies.
Volvo AB is dedicated to offer a wide range of transport and infrastructure solutions to
pioneer sustainable innovations and lead by being upfront and transparent. With the help of
machine learning, a large volume of data can be collected for further analysis to detect
patterns, and by that prediction comprehend future potential. Volvo AB has implemented a
centralized Enterprise Risk Management system, which involves systematic and structured
processes for reporting, reviewing, mitigating, and identifying risks. Risks are categorized
into different areas such as macro and market-related risks, operational risks, climate risks,
people risks, compliance risks, and financial risks (Volvo AB, 2023). Furthermore, Volvo AB
has integrated an external AI tool, Prewave, into its supply chain infrastructure that
systematically searches through worldwide media to identify relevant crises and disruptions
that potentially can be linked to their network of suppliers. Prewave has been provided with
Volvo AB’s supplier names and their addresses which distinguishes potential impacts on the
supply chain operations, and efficiently correlates gathered data with the pertinent supplier
networks (Personal Communication, January 2024).
Consequently, Volvo AB has an obvious interest in using AI tools to achieve supply chain
resilience and become proactive. This research investigates further application areas for AI
7
implementation within Volvo AB’s supply chain and how it can be employed to mitigate
disruptions and risks. The research will further fill the existing literature gaps regarding how
to utilize AI within risk management in an organization within the automotive industry with a
large supplier network design.
1.2. Problem Statement
Despite advancements in supply chain management practices, global supply chains remain
vulnerable to various disruptions, spanning from natural disasters and geopolitical conflicts to
economic fluctuations and pandemics. These disruptions can lead to significant financial
losses, operational disruptions, and complications for businesses operating in complex and
global supply networks. While the importance of supply chain risk management and
resilience is widely recognized, there is a need for further research to understand the
underlying factors influencing supply chain vulnerability, transparency and the effectiveness
of risk mitigation strategies. Given Volvo AB’s presence in the global market, they have
recognized the imperative of mitigating risks and disruptions. Understanding these
fluctuations is necessary for adapting their operations to the demands of the global market.
Additionally, with the emergence of artificial intelligence technologies and tools, there is
growing interest in exploring how AI tools can enhance supply chain resilience by providing
predictive analytics, real-time monitoring, and decision-making capabilities. As a global
operator, Volvo AB is increasingly motivated to seek new innovative tools to use for
navigating potential obstacles in the supply chain.
1.3. Research Questions
Q1: Which of the currently available AI technologies for supply chain management are most
suitable for implementation within Volvo AB?
Q2: What are the strategic challenges faced by Volvo AB in implementing AI technologies
for supply chain management?
1.4. Significance of the Study
The significance of this research lies in its potential to address critical gaps in current
understanding regarding the utilization of artificial intelligence in supply chain management
in the case of disruptions. By investigating the application of AI technologies within supply
chains, the study aims to uncover innovative strategies for mitigating risks and enhancing
adaptability, resilience, and responsiveness. Overall, this research endeavor holds significant
potential for advancing the understanding of how AI can be leveraged to enhance supply
chain management practices in the presence of disruptions. By addressing the research
questions, the study aims to contribute valuable insights that can improve strategic
decision-making and facilitate the development of more robust and resilient supply chain
systems for Volvo AB and which also have the potential for application in other
organizations.
8
1.5. Delimitations
The scope of the thesis may be limited to specific aspects of Volvo AB’s supply chain and
risk management practices due to constraints of resources, and the complexity of the subject
matter. Furthermore, the exclusive focus on only AI technologies and not other potentially
relevant technologies narrows the examinations of broader technological advancements that
might contribute to a more comprehensive view of the study. On the other hand, these
delimitations may also lead to a more in-depth exploration and understanding of AI tools as
described in the problem statement.
1.6. Theoretical Framework
1.6.1. Risk Management
Supply chain risk and supply chain uncertainty are two interrelated terms, yet they entail
distinct concepts. It is suggested by Simangunsong et al. (2012) that while risks typically
implies potential negative outcomes, uncertainty encompasses both positive and negative
possibilities. For instance, risks associated with natural disasters solely lead to adverse
outcomes, whereas uncertainty regarding customer demand can lead to both positive or
negative consequences than initially anticipated. With this in mind, supply chain uncertainty
emerges as a broader term where it also can include supply chain risks (Simangunsong et al.,
2012).
Supply chain uncertainty is considered to be one of the major factors behind the need for
supply chain flexibility. There are three uncertainties which are supply uncertainty, internal
uncertainty, and demand uncertainty. Supply uncertainty refers to the uncertainty of material
availability, material price and alternative sourcing availability. Internal uncertainty is related
to machine availability, quality and processing times, which can be internal and labor
processes. The greater the uncertainty is within the internal processes, the more flexibility is
required. Demand uncertainty can be a form of errors in demand forecast, quantity, prices and
timing (Angkiriwang et al., 2014). Flexibility is a strategy needed for companies when facing
uncertainties in the supply chain, it can either be a reactive strategy or a proactive one. The
reactive strategy addresses the environmental uncertainty which can be internal and external
for the organization, while the proactive flexibility allows organizations to alter the way
uncertainties are managed within the market and to shape customer expectations. When
working with reactive uncertainties, some buffering strategies can be applied. For instance
having safety stock, capacity buffers, supplier backups and working with safety lead times. A
safety stock is one of the most commonly adopted approaches for increasing flexibility in an
environment of fluctuating demand and supply. It reduces the probability of inventory
shortage and increases responsiveness within the organization. Capacity buffer could be
utilized as a substitute or of a complement to safety stock. Furthermore, having supplier
backups will ensure availability and improve flexibility in most situations, though it may
increase costs. Regarding safety lead times, this approach allows companies to enhance the
9
availability of materials, thereby becoming more flexible in meeting demand (Angkiriwang et
al., 2014).
For the proactive uncertainties, redesigning strategies can be applied to increase flexibility,
such as, component commonality, postponement, risk pooling, subcontracting or outsourcing,
flexible supply contracts, setup and lead time reduction, and alternative routing or
transportation mode (Angkiriwang et al., 2014). Component commonality is a product design
strategy that enhances flexibility, efficiency, and responsiveness, particularly for companies
with a wide product assortment or uncertain demand. Postponement, involving the design or
redesign of both manufacturing and administrative processes, can significantly improve
flexibility. Subcontracting or outsourcing aids in mitigating risks related to capacity
utilization, especially with uncertain demand, as flexible supply contracts secure supplier
commitments for significant demand increases, thereby improving supply chain flexibility.
While managing supply chain flexibility can be costly, it results in higher performance
through improved market responsiveness, better resource utilization and increased service
levels. Reducing lead times improves flexibility by enabling companies to better respond to
demand uncertainty (Angkiriwang et al., 2014). This can be achieved through redesigning
procurement processes, prioritizing speed over cost in supplier selection, or developing
suppliers for improved lead time management. Furthermore, reducing setup time is crucial for
enhancing volume and mix flexibility within the production. Shorter setup times allow for the
economic production of small quantities, thus enhancing cost efficiency, flexibility and
responsiveness. Selecting suppliers with short setup times is considered essential for supply
flexibility. Adopting alternative transportation routes or modes enhances flexibility when
sudden orders arise or issues occur on existing routes. The outcome of handling supply chain
flexibility except for being costly, is that it leads to higher performances by improving market
responsiveness, increased resource utilization and increased service level (Angkiriwang et al.,
2014).
Baryannis et al. (2019) further elaborates on the different tiers of risk classifications prevalent
in supply chain risk management. These classifications encompass three primary categories:
environmental risks, network risks, and organizational risks. Environmental risks are external
risks that are connected to the entire supply chain, network risks are from external factors but
do affect the network that the organization operates in, and organization risks are those
internal to the firm. Risk management emphasizes the importance of building resilience and
robustness in the supply chain, as it effectively decreases supply chain vulnerability.
1.6.2. Disruptions in the Supply Chain
The globalization and continual expansion of supply chains have heightened their exposure
and vulnerability to potential disruptions. Such disruptions can cause a negative impact on the
financial performance of the entities involved (Duong Khanh & Chong, 2020). The initial and
fundamental step in risk management is to identify potential risks within the supply chain. A
systematic classification of these risks can aid firms in selecting appropriate risk mitigation
strategies. One proposed classification scheme encompasses factors such as human resources,
10
infrastructure, management processes, external environment, and product characteristics
(Hudnurkar et al., 2017). Furthermore, risks within the supply chain can be classified based
on sources of risks such as uncertainties within processes, supply, demand and control.
Process uncertainty relates to uncertainties regarding internal capacities within the
organization to achieve planned production goals, while supply uncertainty refers to instances
where a supplier fails to fulfill the organization's requirements in terms of timing, quantity,
quality or price. Additionally, demand uncertainty concerns the predictability of demand and
the product variability, while control uncertainty pertains to information flow within the
organization and its capacity to convert orders into production goals and material requests
(Barroso et al., 2008).
Additionally, risks within the supply chain can be classified into seven categories: external
risks, time risks, information risks, financial risks, supply risks, operational risks and demand
risks. External risks include threats from outside the supply chain, caused by economic,
sociopolitical, or geographical factors. For instance, these risks involve economic downturns,
external legal issues, and natural disasters. Time risks refers to delays that could occur within
supply chain processes, while information risks concern complications in information
infrastructure and data leaks. Financial risks involve inflation, currency fluctuations and
stakeholder demands. Supply risks encompass issues related to suppliers, such as unstable
quality, price fluctuations, and supplier bankruptcy. Furthermore, operational risks are caused
by internal issues within the organizations stemming from changes in design and technology
or accidents. Lastly, demand risks refer to variability in demand, high market competition and
customer fragmentation (Ivanov, 2021).
Barroso et al. (2008) proposed a conceptual framework for characterizing various supply
chain disruptions. The framework encompasses dimensions such as time-to-disturbance,
timespan and scope. The first refers to the period between the forecast and the occurrence of
the disturbance. For instance, while a hurricane can be forecasted with a certain degree of
accuracy, earthquakes do not allow any preparation time. The duration of the disturbances,
referred to as timespan, indicates the length of time the organization will be exposed to the
disturbance, thereby affecting its outcome and impact. Additionally, disruptions can have
local or global incidence (scope), meaning they may affect only a specific area within the
organization or extend to broader operations, such as international transportation, resulting in
a global impact on the entire supply chain. Establishing a common vocabulary for risk
identification and assessment among actors within the supply chain is crucial for
standardizing risk mitigation strategies. Given the uniqueness of supply chains, it is
imperative to have a universally accepted classification (Hudnurkar et al., 2017).
Collaboration is considered a crucial factor for enhancing visibility, velocity and flexibility
within the supply chain, resulting in improved resilience. In the presence of disruptions,
organizations often face resource limitations, necessitating adaptations within business
processes. Supply chain partners can consider various adaptations including parametric,
process or structural adjustments. Parametric adaptation involves adjusting critical parameters
within contracts, such as production plans according to the event of disruptions. Process
11
adaptation focuses on the flexibility of capacity within the supply chain, necessitating
partnerships with logistics service providers and the entire supply chain network. Investments
in additional capacity at suppliers or logistics providers can facilitate a more efficient
recovery during capacity disruptions (Duong Khanh & Chong, 2020). Lastly, structural
adaptation involves modifying the supply chain network and deploying assets such as backup
suppliers. Leveraging backup suppliers, alternative transportation plans, schedules and
contingency plans can aid and facilitate in minimizing damage in the event of disruptions.
Additionally, collaboration among partners within the supply chain is considered
advantageous in uncertain environments. Such collaboration enables improved capabilities,
resources, financing and service enhancements for responding to and recovering from
disruptions. Partners are able to share warehouses, vehicles, or labor to mitigate financial
losses and enhance cost control, which can be challenging during disruptions. These
collaborative efforts lead to reduced lead times and more accurate information flows within
the supply chain, facilitating a more efficient and fast recovery (Duong Khanh & Chong,
2020).
In addition to increasing collaboration among supply chain partners, other strategies have
been developed to enhance resilience and mitigate disruptions. These include postponement,
developing strategic stocks, employing a flexible supplier base, a mix of in-house production
and outsourcing, and offering economic supply incentives to increase the number of suppliers
(Marley et al., 2014). Furthermore, actions such as updating supply chain engineering models
and enhancing the ability of supply chain actors to respond to emerging issues are essential
for building resilience. Strategies aimed at increasing flexibility, velocity and developing risk
management systems have been shown to reduce a firm’s vulnerability to potential
disruptions. Interactive complexity has proven to play a crucial role in predicting the
likelihood of disruptions. High interactive complexity may render conventional risk
mitigation approaches ineffective and exacerbate the impact of common disruptions. The
combination of low inventory and interactive complexity results in the fewest disruptions
downstream in the supply chain. However, a higher inventory level may increase the
probability of disruptions when processes within the supply chain are complex (Marley et al.,
2014).
Over the last decades, supply chains have become increasingly complex, which has been
argued to decrease operational performance. Different types of complexities including
vertical, horizontal, and spatial are associated with increased uncertainty, reduced
transparency, and a higher frequency of disruptions. Vertical complexity within organizations
refers to the number of hierarchical levels or tiers. It increases the risk of potential chain
reactions (ripple effect), where disruptions in one tier can cascade down the supply chain,
causing significant disruptions. This complexity also leads to increased uncertainty in the
upstream supply chain, as firms often lack transparency and knowledge about their lower-tier
suppliers, making it challenging to detect and respond to potential disruptions. However,
horizontal complexity is associated with the specialization of competencies and knowledge
within an organization (Bode & Wagner, 2014). In the upstream supply chain, horizontal
complexity is linked to the number of direct suppliers. Multi-sourcing arrangements increases
12
this complexity but has the capability to mitigate the severity of disruptions experienced.
Lastly, spatial complexity refers to the geographical operating location of an organization and
its supply base. This complexity arises from global sourcing practices, resulting in longer and
more vulnerable lead times and reliance on critical infrastructure vulnerable to various risks.
Longer global supply chains are associated with increased risks of disturbances, such as cargo
theft, environmental conditions, and natural disasters. Additionally, managing operations and
communication becomes more challenging due to geographic distance, resulting in
uncertainties, monitoring of costs, and coordination difficulties. Achieving stability within
operations and the supply chain requires firms to understand how the structure of their supply
chains affects the occurrence of disruptions and risks (Bode & Wagner, 2014).
Chopra & Sodhi (2014) proposes that businesses should implement strategies to segment the
supply chain and to treat predictable and unpredictable demand separately. Those strategies
can significantly reduce supply chain vulnerability while improving financial performance.
Through the alignment of supply chain structures with product characteristics and adopting a
differentiated approach of managing demand, companies can enhance their resilience to
disruptions and achieve sustainable growth in today’s business environment. Through
diversifying sources of supply, the impact of disruption risk from a single production facility
is significantly reduced and segmenting the supply chain based upon product volume, variety
and demand uncertainty enhances both profitability and resilience. However, organizations
must find a balance, as excessive redundancy may lead to unjustifiable costs (Chopra &
Sodhi, 2014).
1.6.3. Enterprise Risk Management System
Enterprise Risk Management (ERM) advocates for a comprehensive approach to managing
risks within an organization, emphasizing the importance of aligning risk management with
corporate governance and strategy. Rather than addressing risks individually, ERM seeks to
integrate risk management practices across all levels of the organization. This approach is
underpinned by the belief that managing risks collectively and coherently is more effective
than addressing them individually. Moreover, ERM encourages firms to leverage risk
management to gain competitive advantage, perceiving risks not solely as problems to be
solved but as opportunities for market opportunities (Bromiley et al., 2015). Traditionally,
risk management frameworks have difficulties with effectively addressing risks from multiple
perspectives, underscoring the need for ERM adoption (Razali & Tahir, 2011). However, a
significant challenge facing ERM is the lack of robust tools for facilitating the capture,
analysis, and communication of risk information (Sharma, 2019).
Despite these challenges, organizations increasingly recognize the value of ERM in
streamlining processes and managing human resources. Nevertheless, the absence of
adequate tools to support the implementation of ERM remains a critical issue. AI presents a
promising solution to enhance ERM practices, particularly in terms of incident prediction and
risk assessment automation. AI enables the efficient analysis of vast datasets, providing
insights that contribute to a more comprehensive understanding of risks within the
13
organization. Furthermore, while ERM excels at quantifying and identifying an organization's
exposure to risks, AI offers the potential to identify dependencies and correlations between
risks, thereby enhancing risk management capabilities (Sharma, 2019). However, it is
essential to recognize potential biases that may arise, such as endogeneity, wherein the
adoption of ERM by high-performing entities could falsely associate ERM with improved
firm performance. Additionally, differing managerial perceptions of risks may pose
challenges in aggregating risks effectively within the organization (Bromiley et al., 2015).
Scenario planning is a widely adopted method for companies to develop their long-term
strategies, with a specific focus on identifying and formulating plans on how to manage the
risks. By applying AI techniques to scenario planning, organizations can gain a unique
advantage in foreseeing future scenarios based on observations from news and social media
posts. An AI-driven tool called Scenario Planning Adviser (SPA) has been introduced for
ERM. SPA leverages raw data, including news and social media posts, and interacts with
users to gather observations. Additionally, SPA incorporates domain knowledge from externs
through Mind Maps and automatically generates questionnaires to assess further information.
The tool generates scenarios by clustering multiple quality plans, making them easily
understandable and comprehensible for human users (Sohbrai et al., 2018).
2. Literature Review
This chapter provides a comprehensive literature review on focused application areas of AI in
the Supply chain management, Industry 4.0 technologies, Supply chain risks,
Implementations strategies of AI and existing gaps in the literature. The literature review will
give the reader the necessary theory needed to understand the research topic.
2.1. Application Areas of AI in Supply Chain Management
2.1.1. Forecasting
Demand forecasting plays an important role in supply chain management since demand
uncertainties are known to impact supply chain performance. Therefore, this highlights the
necessity of applying various statistical analysis techniques such as time-series analysis for
demand forecasting (Seyeden & Mafakheri, 2020). It is crucial for planning and procurement
processes, enhancing the responsiveness and efficiency of the supply chain. Traditional
forecasting methods, including statistical techniques, struggle with increasing data
dimensions and volumes, failing to meet the requirement of manufacturing companies. AI
technologies, either alone or combined with statistical methods, significantly improves the
accuracy of demand forecasting within the supply chain. It is crucial to select an appropriate
AI solution to avoid demand distortion (Mediavilla et al., 2022). Additionally, AI can predict
potential shortages that could impact operations, aiding to reduce costs related to logistics,
inventory, and service levels (Kashem et al., 2023).
14
Intelligent forecasting utilizes historical data to adapt and estimate calculations for any
changes of demand in supply chains with the ability of big data analytics techniques where
predictive analytics are primarily used. In comparison to conventional methods, data-driven
techniques can provide more accurate estimations in demand forecasting resulting in
increased efficiency and resilience (Seyeden & Mafakheri, 2020). Organizations have
integrated AI tools across diverse operations of the supply chain, especially in areas to
improve decision-making processes. Some functions that this can be applied to are inventory
management for inventory levels, predicting and forecasting demand, and mitigating any
risks in the supply chain. The predominant usage of AI learning tools in supply chain
contexts is, however, related to forecasting, where for instance, stochastic programming and
fuzzy logic can be applied (Pournader et al. 2021). Fuzzy programming is an applied
technology for managing uncertainties and subjectivity within supply chains, making it a
valuable method for improving its performance (Ganga & Caprinetti, 2011). It enhances the
decision-making process by providing various suggestions and recommendations with
different degrees of satisfaction (Peidro et al., 2010). Moreover, stochastic programming can
assist in reaching supply chain decisions (Govindan & T.C.E, 2018). It commonly utilizes
expected value as the performance measure for finding optimal solutions in the event of
uncertainties (Mo et al., 2010). According to Riahi et al (2021), demand forecasting is the
most promising application for implementing ML and AI.
Browning et al. (2023) argue that forecasting is affected by human cognitive limitations and
biases inherent in judgment, which can cause significant forecasting errors. By utilizing
visual analytics with human augmentation, forecasts can improve the resilience of supply
chains. Furthermore, the employment of AI in forecasting can increase visibility, easen
optimization decision-making processes while maintaining high precision, dependability and
authenticity (Kashem et al., 2023). Other benefits encompass competitive advantage
enhancement, enabling accurate forecasting and effectively managing potential risks (Sharma
et al., 2022). Lastly, Big data analytics is utilized in both forms of supervised and
unsupervised learning approaches in demand forecasting, where the supervised approach
estimates are based on historical data and the unsupervised analyzes the input in order to find
patterns and their correlations (Seydan & Mafakheri, 2020).
2.1.2. Procurement
Procurement can be defined as any purchasing activity, encompassing leasing, buying,
acquiring suppliers, services or construction from external sources (Jahani et al., 2021). The
primary objective of procurement is to reduce costs of buying, optimize procurement
expenses, secure supplies, enhance visibility, and streamline internal processes (Allal-Chérif
et al., 2021). Procurement employs various strategies to address disruptions within the supply
chain. Establishing long-term relationships enhances firm performance by increasing the
likelihood of successful transactions with suppliers. Furthermore, it facilitates efficient
mitigation of potential disruptions through renegotiation. Suppliers tend to address
disruptions more effectively when engaged in long-term relationships or renegotiation with
buyers who offer higher initial purchase prices. Additionally, effective communication also
15
significantly impacts performance. It not only raises purchase prices but also encourages
suppliers to mitigate and address disruptions promptly (Katok & Tan, 2018). Emergency
procurement is widely adopted as a strategy for managing supply chain risks and disruptions.
Under this approach, a source not typically involved in producing the disrupted product
temporarily fulfills the supply requirement during shortages. This strategy proves
advantageous because additional costs are incurred only when shortages arise (He et al.,
2016).
Employing artificial intelligence in procurement holds significant potential for automating
processes and supporting employees. It has greatly impacted the performance and operations
of the procurement department by automating operational tasks and providing data-driven
recommendations to aid decision-making (Meyer & Henke, 2023). The integration of AI into
procurement can confer a competitive advantage to firms by effectively managing vast
volumes of data exchanges between suppliers and buying companies, The potential benefits
of leveraging AI in procurement include the automation of repetitive tasks, streamlining
business processes, facilitating price negotiations, efficiently sourcing suitable suppliers, and
analyzing market trends (Meyer & Henke, 2023). Furthermore, it provides contributions to
improved understanding regarding raw material prices, lead times, and business risks faced
by suppliers in various geographical locations (Allal-Chérif et al., 2021). The integration of
AI enhances efficiency by achieving optimal output at minimal costs, allowing buyers to
focus on tasks delivering higher value within their organizations. AI contributes to an
improvement in the accuracy and comprehensiveness of impact data analysis and the control
over consumption by ensuring the reliability of transactional levels (Chopra, 2019). The
integration of AI in procurement results in transformation of the purchasing process from
operational to strategic (Allal-Chérif et al., 2021).
Moreover, the applications of AI in procurement can be classified into five key areas:
automation and optimization of purchasing processes, supplier selection, predictive
purchasing and decision support, supplier relationship management, and collaborative
management and open innovation. One such application involves spend analysis, where AI
automatically categorizes data originating from enterprise resource planning (ERP) systems,
providing a comprehensive overview of spending for the firm. Another example is utilizing
AI to manage incomplete invoices that cannot be automatically assigned to a functional
account, thereby improving process quality. Additionally, firms can implement AI for
intelligent news feeds, which assess relevant data related to purchased commodity groups,
with machine learning algorithms improving selection accuracy over time, providing
advantages in supply market operations (Meyer & Henke, 2023).
To unlock the full potential of artificial intelligence in procurement, it must be supplied with
diverse data, including supplier data, internal data, public data, and subscription data.
Supplier data involves enabling suppliers to submit information among the actors in the chain
that can be shared across multiple buyers on a permission basis, reducing repetitive data and
promoting scalability for suppliers. Additionally, AI can assess internal data, such as spend
analytics, supplier performance, and CRM system data, providing valuable insights and
16
classifications. Public data, though complex due to varying levels of reliability across
different sites, requires smart procurement processes to ensure the accuracy of alerts and
recommendations. Subscription data, containing information on supplier evaluations,
scorecards, weather services and commodity price indices, can also be utilized (Chopra,
2019). According to industry managers, the most beneficial AI functionalities in procurement
are data consolidation and sorting (Guida et al., 2023). Furthermore, AI excels at assessing
and organizing relevant data from various digital sources. With its contextual training in
procurement market intelligence, AI effectively comprehends and categorizes
procurement-related information, including materials, supplier risks and geographic factors
(Chopra, 2019).
George et al. (2023) argue that ChatGPT (an AI-powered language model) could
revolutionize procurement by optimizing costs, risks, and relationships through its instant
language processing and generation capabilities. Currently, ChatGPT is utilized in
communicating with suppliers and translating documents. It is suggested that ChatGPT could
accurately generate purchase orders tailored to specific procurement circumstances and
environment, support data-driven decision-making, and track order confirmations and
statuses through email. Furthermore, it holds the potential to strengthen and improve
procurement risk management by monitoring supply market news, financial data, and
geopolitical shifts, assessing how external forces may disrupt pricing, availability, and
logistics. ChatGPT can identify high-risk suppliers based on performance history (George et
al., 2023). Other commonly adopted AI tools within procurement include softwares such as
Synertrade Accelerate and SAP Ariba. Those systems offer tools for accounting, document
analysis, and dashboard design. These technologies empower buyers to act proactively,
anticipate issues, and reduce potential risks, making the purchasing process more
collaborative and strategic (Allal-Chérif et al., 2021).
Jaggaer SRM systems provide innovative solutions for supplier relationships, contributing to
the development of shared strategies among stakeholders. The system employs intelligent
machine-learning technologies to assist buyers in identifying key elements, sharing their
vision, establishing partnerships, and mobilizing stakeholders for joint projects. Icertics,
another system, leverages AI to provide suggestions and recommendations for optimizing
contract negotiations, reducing time spent on creating and reviewing contracts (Guida et al.,
2023).
While the adoption of digital technologies in procurement has enhanced the visibility and
control over processes, there are still barriers and challenges to overcome. Organizations
often lack a clear understanding of the advantages of AI in procurement processes and face a
shortage of internal skills to implement these technologies (Guida et al., 2023). One key
reason for unsuccessful implementation is a lack of executive leadership, as without a solid
understanding of procurement strategy and organizational context, the impact of AI
technologies within the organization is limited (Barrad, 2020). Other obstacles to adopting AI
in purchasing include firms’ reluctance to invest in and allocate resources for
implementation, lack of transparency about the requirements of AI implementation, and
17
limited experience with the technology (Meyer & Henke, 2023). Additionally, the resistance
to fully dismantle old systems poses challenges related to data consistency and consolidation
(Guida et al., 2023).
2.1.3. Inventory Planning and Management
AI usage in inventory planning and management utilizes predictive analytics, real-time data
processing and enables lean management practices by reducing waste and emphasizing
process efficiency. AI can address issues and costs related to overstocking or understocking
and hence develop strategies for just-in-time inventory management, leading to overall
decreased inventory costs and increased inventory accuracy (Richey et al. 2023). According
to Riahi et al. (2021) AI tools can advance inventory management through automation. AI
techniques generate reports on demand changes, which will save time on forecasting, mitigate
the necessity for future calculations, ultimately enhancing inventory control and
decision-making processes. Artificial neural networks (ANN) is a common AI-tool used in
the inventory management field. It is one of the machine learning techniques employed for
forecasting purposes in inventory demand, regardless of being affected by a bullwhip effect
in inventory (Sharma et al., 2022). The bullwhip effect, also recognized as demand signal
distortion, stands out as a critical challenge within supply chains. It describes the
phenomenon where even a slight fluctuation in consumer demand can cause a substantial
fluctuation in orders received by upstream suppliers. This impacts the supply chain
negatively, resulting in increased costs like overcapacity and surplus inventory (Yang et al.,
2021).
With the recent developments in machine learning, traditional inventory management systems
are transitioning toward more intelligent solutions. Machine learning is being utilized to
improve overall efficiency and increase the flexibility of inventory management processes,
while reducing total inventory costs and improving storage capacity (Ahmadi et al., 2022).
The intelligent inventory system refers to manual tasks that are transformed to automated
tasks, resulting in time and labor savings, and mitigating potential errors caused by humans.
It captures and analyzes real-data including product availability, location and shipment status.
The key benefits of implementing an intelligent inventory system are the provision of
real-time data updates and monitoring of supply chain disruptions (Bhatti & Bauirzhanovna,
2023).
2.1.4. Supply Chain Network
One of the most effective actions of achieving resilience within the supply chain is to create
networks capable of responding rapidly to changed conditions in the environment. This
requires the importance of being aware and understanding the underlying conditions and
standards within the supply chain, meaning establishing high visibility. Furthermore, the
supply chain must acquire the ability to constantly monitor how the actors can most
effectively respond to and mitigate potential disruptions and uncertainties. In other words, the
supply chain network must establish a risk management culture among the involved actors,
due to risks representing a threat towards its resilience. This culture involves risk analysis,
18
risk assessment and report. Apart from collaboration and the development of a risk
management culture, reengineering of the supply chain is considered an essential factor for
enhancing resilience within the supply chain. Cost optimization and enhancement of
customer service are important and prioritized activities within the supply chain, rather than
improving resilience. In this sense, resilience should be designed and incorporated into the
supply chain structure to minimize the supply chain’s exposure to various sources of
disruptions. This is enhanced through an understanding of the network structure, analysis of
the multi-sourcing supplier environment and applying re-engineering practices to
continuously increase the resilience within the supply chain. Enablers of resilience within the
supply chain encompass allowing for flexible and redundant strategies, understanding of the
supply chain structure and organizational learning (Herrera & Janczewski, 2015).
It is a common perception that risk management decreases the cost efficiency of the
organization. In many instances, reducing the risk of disruptions are associated with higher
investments and costs due to such solutions undermining efforts to improve the cost
efficiency within the supply chain. While streamlining operations results in the mitigation of
daily risks and reducing costs, it also increases the supply chain’s vulnerability and exposure
towards disruptions (Chopra & Sodhi, 2014). To allow for cost efficiency during the
mitigation of risks and disruptions within the supply chain is difficult. Balancing risks
requires centralization of resources, whereas managing disruptions demand decentralization.
The supply chain must overcome challenges to achieve increased resilience and cost
efficiency, which encompass overestimating the probability of disruption thus not promising
long-term advantages. Avoiding and dismissing potential risks are considered less expensive
in the short term and therefore the firm must be willing to accept additional investments and
costs of for instance backup suppliers, even though no disruptions might occur. The second
challenge involves how the resilience within the supply chain is measured and implemented.
The leadership within the organization could encourage and convince the employees about
overestimating risks and justify the additional costs. To build a reliable backup source and
establish higher resilience within the supply chain can be considered a strategic global action
(Chopra & Sodhi, 2014).
A complex supply chain network consisting of multiple tiers of suppliers may be more prone
and vulnerable towards disruptions, as one uncertainty at one tier can cause a ripple effect
throughout the network. Thus, supply chains of simpler structure consisting of fewer tiers
may exhibit greater stability, maintaining steady state in the event of potential disruptions and
uncertainties. The structure of the supply chain network can impact the duration needed for
the organization to recover from a disruption. For instance, a supply chain including multiple
transportation routes and links, may recover faster in comparison with a supply chain
consisting of one route. A well-structured network can allow for efficient flow of information
and resources, facilitate identification of alternate suppliers or transportation routes, resulting
in mitigating the effects of disruption and reducing downtime and costs efficiently. This is
due to a well-structured network's ability to allow for a quick transmission of resources and
information, mitigating the impact of disruptions. (Habibi et al., 2023). To effectively
mitigate disruptions within the supply chain, it is essential to understand various types of
19
disruption propagation, whose influence impacts both the organization and the entire supply
chain. The impact of the disruption varies with the firm’s level of resilience and its position in
the supply chain network. At the network level, the supply chain’s performance is equal to
the integrated performance of individual firms included in the network (Li et al., 2021). By
examining how a disruption affects the performance of the firms, resource allocation can be
optimized across the network and improves the performance of the supply chain. Forward
and backward disruption is considered to have a significant impact on the supply chain
network. Backward disruption propagation refers to a disruption’s effects stemming from the
opposite direction of material flow and forward is identified as disruptions arising from the
regular flow. It is stated that the forward disruption propagation generates more impact on the
supply chain network, while the backward disruption affects the distribution network most
negatively. Separating and differentiating those forms of propagation will support effective
decision-making (Li et al., 2021).
2.2. Industry 4.0 Technologies
2.2.1. Artificial Intelligence
The utilization of Artificial Intelligence (AI) applications has surged with promising
outcomes which likewise have raised considerations for future work and business
management. Organizations are investing in AI solutions by successively integrating AI and
through that also enhancing their supply chain operations (Pournader et al., 2021). AI
technologies and methodologies encompass a wide range of approaches that are within the
scope of AI. These include mathematical optimization, network-based problem
representation, agent-based modeling, any sort of automated reasoning, machine learning,
and big data analytics techniques (Baryannis et al., 2018). The extensive implementation and
widespread use of AI in supply chain management (SCM) is attributed to its capacity to
improve decision-making, lead-time, and general operations by allowing visibility and
transparency through the supply chain. Furthermore, its implementation allows the business
to predict bottlenecks and optimize production planning, thus reducing waste and costs.
Another reason that organizations implement AI-powered software or technologies is to
obtain analytical insights by leveraging predictions made through cognitive computing
techniques which also contributes to improving overall SCM efficiency (Sharma et al., 2021).
Barghava et al (2022) mentions that depending on the volume of data that is fed to the AI
algorithms, the probability of the accuracy and the precision is affected. Larger enterprises
that encompass extensive data resources are most likely to not face issues since AI models
thrive by the increasing scale of industry, whereas small and medium-enterprises may
encounter challenges due to data limitations.
2.2.1.1. Advantages of Artificial Intelligence
There are numerous advantages associated with the implementation of artificial intelligence
(AI) in supply chain management. AI provides and enables end-to-end visibility and
transparency throughout the supply chain, facilitating quick and responsive decision-making.
20
This capability enables organizations to effectively anticipate potential bottlenecks,
streamline production planning, implement smart maintenance, optimize smart service
operations and enhance scheduling. The real-time assessment and processing of information
empower firms to predict seasonal fluctuations and reduce the bullwhip effect through
improved resource planning and demand-driven manufacturing (Sharma et al., 2022).
Artificial intelligence holds the capacity to enhance dynamic abilities in sensing,
transforming and seizing, thereby reducing the risk of current disruptions occurring and
proactively avoiding future issues (Belhadi et al., 2021). Additionally, other advantages of
incorporating AI in the supply chain include accurate forecasting capabilities, ensuring safe
working conditions, improving overall quality and offering powerful optimization capabilities
(Sharma et al., 2022).
The utilization of artificial intelligence in the supply chain is also perceived as a contributor
to the emergence of new supply chain structures, indirectly enhancing collective behavior and
enabling improved responsiveness and efficiency. AI adds value to the supply chain by
facilitating accurate forecasting and fully adopting autonomous planning techniques.
Furthermore, AI contributes to knowledge creation and distribution across the chain,
providing firms with a sustainable competitive advantage (Ghetto, 2021).
Riahi et al. (2021) argue that AI adoption within the supply chain leads to increased
performance, lower costs, reduced losses and comprehensive changes that render the supply
chain adaptable, agile and resilient. AI-driven innovations have a high potential to enable
firms operating in highly dynamic environments to enhance, or at least maintain, their current
performance levels within the supply chain. This is achievable due to AI’s ability to learn
from data, adapt decision-making processes, promote supply chain innovation and respond
swiftly during disruptive events (Belhadi et al., 2021). It can be concluded that artificial
intelligence aids in forecasting demand, optimization, promoting, delivery and pricing which
contribute to an increased competitive advantage for the firm (Dash et al., 2019).
2.1.1.2. Challenges with Artificial Intelligence
Despite immense advantages, artificial intelligence consists of an array of hurdles and
challenges that the organization must manage and mitigate. These challenges are spanning
over ethical considerations, data privacy concerns, transparency and visibility. Examining
these challenges is crucial for an organization, as it promotes the development of a supply
chain management system that not only incorporates advanced technology but is also
efficient, robust and environmentally sustainable (Richey et al., 2023).
It is vital to recognize that artificial intelligence lacks free will and relies heavily on computer
software (Min, 2010). Implementing AI requires generous amounts of data for optimal
functionality. Acquiring and generating such data can be considered resource-intensive
(Richey et al., 2023). For successful implementation and integration of AI within the supply
chain, organizations must ensure the availability of the right expertise and competencies to
employ these new techniques, along with access to suitable data (Lynn et al., 2019). Incorrect
21
AI predictions and results can cause significant economic consequences, such as overstocking
or stockouts (Richey et al., 2023). Additionally, the perceived high investment associated
with implementation of AI makes organizations hesitant to invest and face the risk of scaling
(Ganesh et al., 2022; Dogru & Keskin, 2020).
Ethical and safety concerns surround the use of artificial intelligence in the supply chain. It
can be perceived that AI operates in a legal gray area characterized by unclear regulations,
raising issues related to accountability and trust. Inadequate and insufficient regulations foster
an asymmetric field, where profit-driven organizations are motivated to expedite artificial
intelligence development, potentially neglecting crucial tests to protect consumers from harm
and biases. Furthermore, determining a legal entity accountable on the provider’s end for
consumer harm caused by automated technology becomes challenging (Dogru & Keskin,
2020). Ethical concerns encompass biased behavior in AI systems, manifesting as gender,
racial or age biases in areas such as recruitment, credit scoring and law enforcement. For
logistics and supply chain management, these biases can result in unjust prioritization,
perpetuating unequal opportunities and systemic discrimination (Richey et al., 2023).
Additionally, the increasing replacement of labor by these technologies could create a
significant “displacement effect” in the economy unless balanced by a “reinstatement effect”
(i.e. the creation of new tasks for labor), leading to sustainable productivity gains and
balanced economic growth (Dogru & Keskin, 2020).
The safety considerations and issues associated with incorporating artificial intelligence into
the supply chain revolve around the resilience of cyber systems and the protection of data
privacy. As the demand for personalized and customized services and products continues to
increase, supply chains must manage substantial volumes of personally identifiable
information, encompassing both structured data like geographic location and names, as well
as unstructured data such as videos, tweets and pictures. Ensuring proper data storage and
security demands significant investments in information technologies. Typically, small
technology firms, contracted by supply chains, collect data that is then stored and processed
in data warehouses owned by larger technology corporations. The complex layers of
subcontracting creates an unclear and blurry distinction between the data owner, collector,
controller and processor, thereby exposing consumers to potential vulnerabilities regarding
data privacy. This subcontracting model also heightens the risk of data breaches and the
unauthorized use of personal data for purposes unrelated to the original business intent
(Dogru & Keskin, 2020).
Barriers to the implementation and utilization of artificial intelligence have been identified in
the literature. These barriers encompass challenges within the nature of AI, organizational
structures and operationalization. Organizational barriers include a lack of general
understanding of artificial intelligence, unclear project goals involving AI, ambiguous
ownership boundaries, and misaligned strategies among various supply chain and
management actors. Additionally, successful implementation relies on top management
support, which cannot be achieved in the absence of trust in AI, misidentification of problems
for AI to solve, and a lack of commitment to AI initiatives (Shirvastav, 2022). Overcoming
22
these challenges requires effective change management, emphasizing effective organization
and development practices (Hangl et al., 2022).
Firms must acknowledge the roles of their supply chain partners and organizational processes
when developing AI. The introduction of new technology may face internal resistance,
necessitating effective management, with crucial support from top management. Successful
AI implementation hinges on human acceptance of these techniques, ensuring full
transparency, security, privacy adherence to technical standards, and the robustness of
solutions, including data and documentation (Hangl et al., 2022).
2.2.1.3. Artificial Intelligence of Things (AIoT)
The term “Artificial Intelligence of Things'' refers to the integration of artificial intelligence
and IoT. The purpose of this integration is to enhance human-machine learning, operations in
the IoT field and big data analytics. Currently, AIoT is utilized as a digital technology for
data acquisition, processing, and analysis in manufacturing, aiming to learn, predict
decisions, and provide strategic actions. It establishes intelligent, interconnected systems
where AI functions as the brain for IoT devices. IoT assesses and transmits data from various
sources to enhance the AI’s learning process (Matin et al., 2023). AI has the capability to
gather insights from the assessed data, offering guidance for intelligent actions that enable
business to make informed decisions. The primary advantages of adopting AIoT include
making smart business decisions, increasing operational efficiency, eliminating irrelevant
data, and making accurate predictions about customer behavior (Aliahmadi et al., 2022).
Various application areas exist for AIoT. It possesses the capability to enhance industries by
making them smarter through monitoring processes, detecting potential faults, and preventing
downtime. Additionally, AIoT facilitates intelligent data analysis to support informed
decision-making and real-time data assessment from production lines within the
manufacturing function. It extends its impact to smart transportation by encompassing traffic
participants, infrastructure and industry applications like smart connected logistics. An
illustrative instance is a self-driving car, empowered by AI with perceiving, learning,
reasoning and behaving abilities. Furthermore, AIoT finds application in establishing smart
security within businesses to ensure protection in both the physical world and cyberspace. A
notable feature is human-centric perceiving, recognizing individual identities and analyzing
behaviors to prevent illegal activities (Zhang & Tao, 2021).
Despite its value-generating applications, AIoT presents significant challenges. The intensive
computation it requires poses a computational scheduling challenge, particularly when
scheduling computations across various resources.This necessitates consideration of factors
such as data type, volume, processing latency and performance accuracy. Moreover, the
massive assessment and exploitation of data leads to concerns about data monopoly, wherein
vast prosperity data is protected by established interests, creating a barrier to market entry for
new competitors and posing a threat to free market competition (Zhang & Tao, 2021).
23
2.2.1.4. Explainable Artificial Intelligence (XAI)
Explainable Artificial Intelligence (XAI) provides explanations for its produced outcomes,
facilitating a deeper understanding of its suggested decisions. These explanations enable
stakeholders to assess the fairness, trustworthiness, and reliability of the results. The primary
objective of XAI is to enable organizations to comprehend, trust and effectively utilize the
outcomes generated by AI technologies (Nimmy et al., 2022). It is grounded in three core
principles: transparency, interpretability and explainability (Kangra & Singh, 2022).
Moreover, XAI has introduced a variety of tools and frameworks that aid in comprehending
and interpreting predictions generated by machine learning models, thereby fostering the
development of more understandable and inclusive AI systems. In the domain of supplier
selection, XAI can enhance procedures by employing transparent and comprehensive
methods (Kumar & Kumar, 2023). XAI is commonly utilized to clarify outcomes and results
in decision-making processes. It is essential that decision-makers possess awareness and
understanding of underlying reasons behind AI suggestions; for example, a maintenance
engineer must understand the abnormal behavior and phenomena detected by AI before
implementing suggested repair recommendations (Xu et al., 2019).
Among the identified advantages of employing XAI are improved explainability,
transparency, expedited adoption, enhanced debugging and facilitation of audits to meet
regulatory requirements. Improved explainability and transparency provides organizations
with a deeper understanding of AI models and their behavior under various conditions.
Through an explanation interface, humans can grasp how AI models reach a specific
outcome. As organizations enhance their understanding, they can entrust AI systems with
more critical decisions, leading to faster adoption. XAI can also be employed to diagnose
issues and assist developers in situations where systems start to behave abnormally (Kangra
& Singh, 2022).
XAI is utilized across various domains, such as threat detection and prioritization, where it
improves detection accuracy by adapting to dynamic factors like changing user behaviors.
Nonetheless, a notable hurdle arises from the opaque characteristics of certain methods,
which obstruct a comprehensive understanding of their operations and inference mechanisms.
Additionally, XAI plays a crucial role within guarding against adversarial machine learning,
swiftly identifying and addressing manipulations to enhance confidence and trust in the
integrity of machine learning outcomes (Longo et al., 2020). The primary success factor for
XAI lies in establishing an effective human-AI interface to facilitate interaction between
parties. This involves exploring both the explainability aspect, referring to explanations
produced by machines, and the human side involving human understanding and
interpretation. In an ideal scenario, machine explanations and human understanding would
align, but in reality disparities do exist. Human models, such as problem-solving models,
often rely on casuality, which poses challenges as current machine learning techniques
primarily focus on correlation (Kangra & Singh, 2022).
24
Various challenges and barriers are associated with XAI. Technical challenges include
evaluating the quality of explanations, which partly depends on the receiver’s interpretation
and understanding. Despite its aim to provide explanations and expose the inner workings of
learning techniques and model inferences, it can be difficult for humans to interpret them.
Explanations serve as an intermediate layer requiring expertise, context, and common-sense
characterization for appropriate human interpretation and decision-making. Legal challenges
also exist regarding the usage and implementation of XAI, particularly concerning sensitive
information management (Longo et al., 2020). Additionally, other obstacles faced by XAI
include complexity, irrationality, context dependency, confidentiality concerns, and lack of
expertise. AI algorithms may produce biased or irrelevant conclusions, raising fairness
concerns. Results must be contextualized based upon the data provided to AI systems. Many
individuals within organizations may lack the necessary competencies to comprehend
explanations generated by AI. Moreover, XAI’s results may vary based on the context and
concerns about data security that arise due to the vast amounts of data managed by the
system. Despite XAI algorithms being perceived as simple to understand, deploying them can
become complex. While it may provide and generate more efficient and understandable
algorithms for inexperienced users (Kangra & Singh, 2022), concerns could arise regarding
its usability for experienced users.
Disasters have emerged as a significant global challenge, posing threats to infrastructure and
economies, underscoring the imperative of assessing and mitigating risks. Addressing this
necessitates collaboration among stakeholders, including governments and non-governmental
organizations. Managing the risks of disasters entails a complex decision-making process
reliant on accurate, reliable, and timely information, which XAI is equipped to facilitate and
support. AI within disaster risk management offers promising opportunities for enhanced
efficiency, precision, and effectiveness in responding to disasters and emergencies. For
instance, it can be utilized to analyze geospatial data, aiding authorities in assessing the extent
of damage following a disaster. Moreover, it contributes to long-term resilience by
identifying vulnerable areas and offering sustainable solutions. For instance, AI could be
deployed within climate modeling to identify potential risk areas for sea-level rise, informing
coastal zone management to bolster preparedness and resilience. Overall, XAI provides
insights into the underlying factors driving disaster risks and the efficacy of various disaster
risk management strategies, fostering increased trust among stakeholders in AI-based
decision support systems, thereby enhancing both their acceptability and adoption
(Ghaffarian et al., 2023).
Disaster risk management (DRM) involves four primary phases: prevention-mitigation,
preparedness, response, and recovery. Achieving effective DRM requires integrating all these
stages, emphasizing the importance of proactive measures such as prevention and mitigation.
XAI has the potential to contribute across the DRM phases. During the prevention and
mitigation phase, it has the capability to automate data processing for assessing impact and
damages, providing comprehensive explanations about the utilized features. Additionally,
XAI can develop predictive models that incorporate diverse data sources, such as climate
25
data, facilitating the visualization and mapping of damages linked to various types of disaster
risks (Ghaffarian et al., 2023).
2.2.2. Machine Learning
In recent years, AI research related to Machine Learning (ML) has expanded. Techniques
within ML enable computer programs to learn from input data which can operate with or
without given information about the expected output. This means that ML approaches
develop systems capable of predicting outcomes and learning patterns solely from the
received input data. Moreover, ML functions within the automation of decision-making
regarding supply chain risk management. This can be achieved by unsupervised learning
algorithms that will assist in the process of identifying and understanding various risks within
the supply chain by recognizing and extracting patterns. The algorithm can undergo training
to recognize patterns associated with risks by providing examples of patterns and employ
learning-based techniques for classification and prediction which enables the assessment of
the levels of risk present (Baryannis et al., 2018). Supervised ML generates predictions on
desired outputs by training a machine that uses a pre-existing data set, whereas reinforcement
learning resembles unsupervised learning but yet differs by receiving feedback post-task
completion and enhancing performance through trials and errors (Pournader et al., 2021).
According to Lynn et al. (2019), the output generated from ML needs to be interpreted and
applied in a specific context by human decision-makers, to ensure that the insights provided
are applied accordingly.
ML can be used in different parts of the supply chain where decision-making is involved.
Except for risk identification, it can also be applicable in trend forecasting and inventory
accuracy by using different ML algorithms (Sharma et al., 2021). By employing ML,
organizations benefit since it enables them to gain insights into different areas in the supply
chain, due to its ability to analyze an extensive amount of data and information and generate
structured results. ML’s ability to analyze an extensive and large amount of data and generate
structured data, leads to increased efficiency by the new knowledge provided. ML can
therefore assist in decision-making and increase resilience in the supply chain (Riahi et al.,
2021).
2.2.3. Cloud Computing
Cloud computing, as defined by Kochan et al., (2014), refers to the provision of
reconfigurable resources such as software, infrastructure, or platforms, facilitated through
connectivity. This model serves various organizational needs and operations and as an
Information Communication Technology (ICT) sourcing method. It grants convenient,
on-demand network access to a shared pool of virtualized resources, characterized by features
such as on-demand self-service, broad network access, resource pooling, rapid elasticity, and
measured services (Herrera & Janczewski, 2015). The adoption of cloud computing not only
enhances structural flexibility but also improves responsiveness, thus fostering competitive
advantages for firms (Gammelgaard & Nowkicka, 2023).
26
Integrating cloud computing into the supply chain holds significant potential for enhancing
resilience. By swiftly identifying and mitigating disruptions, cloud technology aids
downstream suppliers in preparing contingency measures, thus minimizing the adverse
impacts of disruptions on supply chain performance (Chen & Chang, 2021). Stakeholders
within the supply chain must define resilience requirements based on organizational drivers,
risk tolerances, and objectives. Utilizing cloud technology facilitates the prevention and
management of disruptions, fostering operational resilience and supply chain collaboration
(Herrera & Janczewski, 2015).
Cloud computing’s shared pool of IT tools enables cost reduction, scalability maximization,
and rapid development, thereby enhancing agility, flexibility, and efficiency within the supply
chain (Giannakis et al., 2019). Coordination mechanisms within cloud computing, including
protection, response, and adaptation, mitigate operational risks and potential supply chain
damage. It is essential to understand cloud computing’s architecture, characteristics, and
relationships to establish effective protection mechanisms. These mechanisms should
encompass situational awareness, vulnerability identification, response coordination, and
knowledge sharing to maintain resilience (Herrera & Janczewski, 2015). However, the
adoption of cloud computing also introduces complexities and challenges, particularly
concerning cyber-related disruptions. Cybercriminals exploit vulnerabilities within supply
chains, heightening risks associated with cloud adoption. The “Supply Chain Cloud Risk
Assessment” model assists IT managers in improving defenses against cyber threats by
systematically analyzing potential risks (Chen & Chang, 2021). Despite its advantages, cloud
computing operates in a highly dynamic environment, contributing to operational risks
(Herrera & Janczewski, 2015).
2.2.4. Internet of Things (IoT)
The Internet of Things (IoT) comprises sensors and devices that collect and process data from
complex environments, transmitting it to cloud centers via the internet for connectivity and
perception. It represents a global intelligent network that facilitates interactions by linking
numerous devices with the ability to perceive, compute, execute and communicate over the
internet. IoT enables information exchange among various smart devices, users, and data
centers (Zhang & Tao, 2021). It has the capability of discerning relationships between people,
vehicles and other entities, as well as recognizing patterns too complex for the human mind to
identify (Greengard, 2021).
There are several application areas for the IoT within supply chain management, including
inventory management, asset tracking, predictive maintenance, route optimization and waste
reduction. In each application area, the IoT aims to enhance efficiency, reduce costs, ensure
product quality and promote sustainability within the supply chain. It represents a disruptive
force seeking to reshape supply chain operations. The integration of IoT technologies enables
real-time tracking and data-driven decision-making (Sallam et al., 2023). Furthermore, the
IoT has various impacts on supply chains, including providing real-time temporal and spatial
27
information on product flows, aiding in the detection and resolution of inconsistencies and
safety crises (Ben-Daya et al., 2022).
One of the primary challenges in supply chain management is matching supply with demand
due to uncertainty and a lack of synchronization among partners. The IoT can bridge this gap
by providing rich information to better understand customers’ demands, thus reducing
uncertainties and facilitating demand-supply synchronization (Ben-Daya et al., 2022).
Additionally, the IoT’s real-time visibility allows organizations to adapt quickly to changes in
demand, address stockouts or overstocking, and optimize inventory levels. By analyzing
historical data and real-time trends, the IoT can improve demand predictions, enhancing
accuracy, reducing costs, and improving forecasting. This transparency facilitates real-time
decision making (Sallam et al., 2023).
Alongside its potential to improve the SCM performance, the IoT presents some complexities
and challenges. These complexities include data security and privacy challenges, as well as
interoperability issues. The absence of universal standards obstruct communication between
IoT devices, and without standardized data formats and interfaces, integrating IoT devices
becomes labor-intensive and costly. Legacy systems in many supply chains, including older
machinery and infrastructure, do not support new technologies like IoT, making integration
resource-intensive, complex and expensive. Successful implementation and integration of the
IoT requires the right technical expertise and competencies (Sallam et al., 2023). Moreover,
ensuring the safety and information security of IoT systems is a challenging and a difficult
task to complete (Kopetz & Steiner, 2022). Additionally, transmitting vast amounts of data
from various environments and making intelligent decisions based upon the information
received requires substantial bandwidth and cloud processing, further adding to the
complexity (Zhang & Tao, 2021).
2.2.5. Artificial Neural Networks (ANN)
Artificial Neural Network (ANN) is a type of ML and is in the context of supply chain
management applied for demand forecasting, anomaly detection and inventory optimization
(Soori et al., 2023). It can be utilized to compute anticipated costs and production loss risks.
ANN is equipped with input data encompassing scenarios featuring production times,
quantities and capacities while generating output in the form of corresponding cost estimates.
Through training on this dataset, ANNs acquire the capability to discern correlations between
input and output, consequently yielding more reliable results. ANN also eases the process of
selecting suppliers by evaluating the suppliers’ reliability and providing flexibility, which
also means enhanced decision-making in organizations (Atwani et al. 2020). According to
Toorajipour et al. (2021), ANN is the most common AI technique applied in supply chain
management, tasked with identifying, comprehending and predicting patterns from a vast
dataset that humans can not find. Its wide functionality and capability to solve intensive data,
enables applications in different areas of the supply chain, such as in sales and production
forecasting, demand management and supplier selection. ANN is effective both when applied
by itself or in combination with other AI techniques, when in combination it enhances
28
accuracy and improves performance (Toorajipour et al., 2021). Additionally, ANN is
considered a promising tool for analyzing and evaluating supplier performance metrics
encompassing quality, delivery times and pricing. Through leveraging ANN within risk
management practices within the supply chain management, organizations can enhance their
ability to identify and proactively address risks and disruptions (Soori et al., 2023).
ANN have demonstrated significant promise in the realm of risk management within supply
chains. Leveraging their capacity to identify patterns and correlations, ANN can play a
pivotal role in assessing the likelihood and impacts of potential disruptions, encompassing
natural disasters, supplier issues and market volatility. By employing ANN in risk
management, supply chains can enhance resilience through proactive risk mitigation
strategies (Soori et al., 2023). Supply chains are typically reliant on data and information that
are often incomplete and inaccurate, originating from diverse sources. ANN’s ability to
handle incomplete data and mitigate uncertainties can be useful in this context. By examining
patterns within the data, ANN can effectively discern inconsistencies from regularities. In a
study investigating the fulfillment status and duration of upcoming orders using ANN, it was
found that the recognition rate exceeded 99,5% underscoring the efficacy of ANN in
predicting potential disruptions within the supply chain (Silva et al., 2017).
There have been some case studies where ANN has been proposed to be used in combination
with a gray model applied in a case study regarding predicting demand of transportation
disruptions. Furthermore, Seyedan & Mafakheri (2020) examined the application of ANN in
forecasting spare parts within the aircraft industry, with fluctuations in the demand, high
diversity of needs and a variety of customers. Another case study presented by Oliveira et al.
(2013) acknowledges ANNs ability in forecasting stock prices within the finance market with
higher accuracy compared to conventional methods. This is further stated through another
case study where ANN in combination with genetic algorithms developed a neural-network
forecasting model which resulted in remarkable accuracy in estimating forecast (Seyedan &
Mafakheri, 2020). Genetic algorithms are generally search based algorithms that are
commonly utilized for machine learning and problem optimization (Lambora et al., 2019).
Whereas Lim et al (2022) accentuated the combination of ANN and Structural equation
modeling (SEM) considering the results of a case study where it gave a deeper insight and
resulted in a more accurate depiction of underlying connections and patterns. Structural
equation modeling is a multivariate technique utilized in investigations aimed to test and
evaluate direct and indirect effects on pre-assumed causal relationships (Fan et al., 2016).
One of the key advantages of utilizing ANN as an analytical approach is its ability to
effectively master both linear and non-linear correlation between independent and dependent
variables (Lim et al., 2022).
Ahmafori et al. (2017) conducted research exploring the implementation of ANN for
predicting completion times in a supply chain system for an audio speaker manufacturer. The
study revealed that ANN outperformed regression methods in predicting completion times,
highlighting its ability in leveraging historical data to forecast future occurrences.
Furthermore, ANN assigns weights to related variables, with variables of greater weight
29
signifying greater influence. Another study aimed at enhancing the performance and
efficiency of SCM through the adoption of ANN. The study presented a notable improvement
in the supply chain network’s performance, with an increase of two or three percent upon
implementing the ANN algorithm. However, it was noted that ANN is most effective when
applied to problems with simple input and output data, as larger datasets can introduce
complexity and interpretability difficulties. Additionally, an excessive number of hidden
nodes in the ANN model may lead to suboptimal results (Praveen et al., 2019). ANN
encompasses various data problems regarding imbalanced data, incomplete data, high
dimensionality and limited data accessible for learning (Prieto et al., 2016).
2.2.6. Agent-Based System
An agent-based AI system is a computational framework designed to replicate the behaviors
and interactions of autonomous agents, whether acting collectively or independently. It
accomplishes this by factoring potential influences that could impact the system overall.
Agents are entities with the ability to perceive their environment and respond autonomously
and proactively to address particular challenges. For instance, agent-based models have been
employed within application areas including supply chain planning, simulation of systems
and analysis of complex behavior within the supply chain (Toorajipour et al., 2021).
An environment characterized with multiple agents communicating, interacting and
coordinating their actions and behavior, requires a more coherent form of agent-based
technology known as Multi-Agent Systems (MAS). MAS comprises multiple interacting
agents, which can be either homogeneous (agents with similar capabilities) or heterogenous
(involving different individuals and objectives) (Kaur & Kaur, 2022). The agents within the
system are symbolizing and representing actors within the supply chain such as suppliers,
customers, manufacturers, and retailers. The system adapts their behavior to anticipate,
respond, enhance flexibility, and recover from both predicted and unexpected disruptions
when simulating various scenarios within the supply chain (Kassa et al., 2023). MAS is
adopted to manage complex problems that are considered too challenging and difficult for
individual agents to solve (Kaur & Kaur, 2022). Through leveraging on the agent-systems
learning capability, it can provide proactive and autonomous responses for the participating
agents to mitigate and rectify disruptions in real-time (Kassa et al., 2023). This allows for
improved coordination, which results in the optimization of time and resource usage. As a
result, MAS are preferred for providing solutions for scenarios where decision-making is
constrained by limited resources and timeframes (Kaur & Kaur, 2022).
Moreover, agent-based systems are utilized to devise various strategies and actions for
managing disruptions and to evaluate their impact on supply chain performance. Equipped
with a database containing historical data on past supply risk events, they can simulate
potential disruption scenarios, including the data on occurrence and duration, facilitating
proactive planning. Additionally, these systems are employed to assess different disruption
management strategies, streamlining the decision-making process (Blos et al., 2015).
Agent-based systems also constitute a crucial role in disaster management by formulating
30
efficient plans and enhancing decision-making in crisis situations (Won Bae et al., 2018).
Multi-Agent Systems are particularly effective in disaster management, enabling enhanced
collective, cooperative, and collaborative planning among involved agents at a large scale.
MAS further facilitates in mitigating and managing uncertainty and conflicting information
during disasters, managing it more effectively (Kaur & Kaur, 2022).
2.2.7. Blockchain
Blockchain technology holds the capability to establish an immutable, distributed, available,
secure and publicly accessible repository of data (Maesa et al., 2019). It is a distributed ledger
that records and shares all transactions occurring within the supply chain. It comprises
multiple nodes and provides an unaltered record of transactions (Azzi et al., 2019). This
technology enhances trust among actors within the supply chain regarding the exchange of
information and data. Within the blockchain, systems are able to communicate directly with
each other (Longo et al., 2019). It encompasses features such as decentralized structure,
distributed nodes, consensus algorithms, and storage mechanisms. Blockchain ensures the
confidentiality, integrity, and availability of all transactions and data within the supply chain.
Other characteristics include autonomy (where every node in the blockchain can access,
transfer, store and update data independently) and openness (providing access to everyone
within the network) (Dutta et al., 2020). A blockchain is not designed to store vast amounts
of data and cannot replace traditional databases. The proposed solution of this technology is
rather to store the original data in an off-chain storage (Longo et al., 2019). In the context of
supply chain management, blockchain technology could be utilized to improve the
traceability of information, processes and financial flows (Chang & Cheng, 2020).
The advantages obtained with the implementation of blockchain includes transparent and
accurate end-to-end tracking, increasing trust and visibility among actors, reducing
administration costs and issues regarding fraud and counterfeit (Azzi et al., 2019). Despite
those advantages, blockchain technology involves various difficulties to manage. The
regulation and laws surrounding blockchain are considered unclear, which could potentially
cause misinterpretations and confusion among stakeholders. Additionally, the operation and
implementation costs of such a system are expensive. Other difficulties involve
interoperability, transition and integration of people, processes and technology. The
difficulties regarding interoperability tend to lead to standardization issues and the unclear
regulations often prevent innovation and developments. The integration of blockchain
involves a considerable change within the supply chain, requiring that all stakeholders are
involved and approve of the new way of thinking and operating. There might be conflicting
objectives for different stakeholders and a lack of awareness will hamper acceptance within
the supply chain (Dutta et al., 2020). Blockchain also faces barriers regarding technical issues
involving transaction throughput, scalability, security, power consumption and privacy
(Chang & Chen, 2020).
31
2.2.7.1 Blockchain in Risk Management
Blockchain technology constitutes a vital role in the risk management process. It holds the
potential for managing and mitigating supply chain risks due to its ability to prevent security
breaches while strengthening supply chain connectivity (Min, 2019). This is achieved
through its capacity to improve security within intelligent transport systems, enhance
implemented industry 4.0 technologies, resulting in improved security for smart objects and
machines. Additionally, the establishment of smart contracts by blockchain will contribute to
improved responsiveness, lead-time reduction, increased visibility, transparency and trust
within the supply chain (Queiroz et al., 2019). Smart contracts improve the visibility within
the supply chain and reduce the need for intermediaries (Dutta et al., 2020). It is considered
an essential application within blockchain and it automates transactions of assets when a
specific condition is satisfied, allowing consumers and producers to trade without
intermediaries (Queiroz et al., 2019). The supply chain could improve its operational
efficiency and mitigate risks through the application of blockchain, leveraging its
characteristics of transparency, decentralization, traceability, and cryptographic security.
Furthermore, blockchain emphasizes proactive risk prevention, including the detection and
prevention of tangible or intangible risks. Risks related to sourcing, involving unclear supply
sources, credit failures, and contract fraud, can be mitigated through blockchain. For instance,
blockchain can establish a complete procurement history that ensures the integrity and safety
of material sources and enhances trust among all stakeholders within the supply chain. The
adoption of smart contracts reduces the risk of fraud in the procurement process (Lai et al.,
2021). According to Hu & Ghadimi (2022) there are four main ways blockchain can mitigate
risks within the supply chain. Through distributed records, the risk of information asymmetry
can be mitigated, data traceability improves enterprise and customer supervision, and
decentralization increases the efficiency of clearing and settlement within the organization.
Lastly, smart contracts resolve operational risks related to cyberattacks, counterfeiting,
miscommunication and credit failures.
Blockchain does not possess the ability of mitigating process and control risks. The value of
implementing blockchain within supply chain management stems from the improvement of
information-sharing and processing capabilities among several organizations or business
units. The technology mitigates risks related to supply and demand, including behavioral
uncertainties, poor information security, fraud and counterfeit risks, data loss and human
errors, operational risks, information asymmetries and transactional risks. The upstream of
the supply chain utilizes blockchain to ensure inputs are derived from sustainable sources,
while the downstream employs the technology to obtain necessary information regarding the
origin, custody, integrity, and authenticity of goods to enable informed decisions and reduce
risk perception (Alkhudary et al., 2022). Furthermore, blockchain enhances the resilience
within the supply chain through improved collaboration among stakeholders strengthened by
transparency and trust. Blockchain allows for full transparency in the supply chain, which
mitigates risks related to contractual agreement. Additionally, the real-time availability of
information enhances resilience still further. The sharing and exchange of data regarding both
risks and information might transform supply chain risk management into becoming more
32
proactive (Lohmer et al., 2020). The real-time availability of data further mitigates
disruptions by ensuring smooth and secure communication among the stakeholders.
Furthermore, potential false information can be detected through blockchain due to its
transparency (Alkhudary et al., 2022).
2.2.8. Summary of Industry 4.0 Technologies
This table displays the various advantages and disadvantages with the respective industry 4.0
technology.
Industry 4.0 Advantages Disadvantages
Technologies
Artificial
Intelligence ● End-to-end visibility and ● Ethical
transparency considerations.
● Quick and responsive ● Data privacy
decision-making concerns.
● Streamlining production ● Dependency on
planning computer software.
● Anticipating potential ● Resource-intensive
bottlenecks data acquisition and
● Optimizing smart service generation.
operations ● Need for expertise
● Real-time assessment and and competencies.
processing of information. ● Economic
● Accurate forecast capabilities. consequences of
● Proactively avoiding future incorrect predictions.
issues. ● Perceived high
● Improving collective behavior. investment.
● Increasing performance. ● Legal gray area and
● Reducing losses and lowering unclear regulations.
costs. ● Accountability and
trust issues.
● Biased behavior in
AI systems.
● Safety considerations
regarding cyber
threats.
● Organizational
barriers.
● Unclear project goals
for the AI.
33
● Internal resistance to
new technology.
Artificial
Intelligence of ● Enhancing human-machine ● Intensive
Things (AIoT) learning. computational
● Improving operations in the requirements.
IoT field. ● Computationalscheduling
● Enhancing big data analytics. challenge.
● Making smart business ● Consideration of
decisions. factors such as data
● Increasing operational type, volume,
efficiency. processing latency
● Eliminating irrelevant data. and performance
● Monitoring processes and accuracy.● Concerns about data
detecting potential faults. monopoly.
● Supporting informed decision ● Barrier to market
making. entry for new
● Accurate predictions about competitors.
consumer behavior. ● Threat to free market
competition
Explainable
Artificial ● Provides explanations for ● Technical challenges
Intelligence produced outcomes. in evaluating the
(XAI) ● Facilitates deeper quality of the
understanding of suggested explanations.
decisions. ● Difficulty in human
● Enables stakeholders to assess interpretation of
fairness, trustworthiness and explanations.
reliability. ● Legal challenges
● Enhances procedures in regarding usage and
supplier selection. implementation.
● Improves explainability and ● Complexity,
transparency. irrationality and
● Facilitates audits to meet context dependency.
regulatory requirements. ● Confidentiality
● Improves detection accuracy in concerns and lack of
threat detection and expertise.
prioritization. ● Variability in results
● Automates data processing for based on context.
assessing impact and damages.
34
● Contributes to effective DRM. ● Concerns about data
security.
● Complexity in
deployment and
usability for
experienced users.
● Biased or irrelevant
conclusions, raising
fairness concerns.
Machine learning
● Enables computer programs to ● Output from ML
learn from input data. needs interpretation
● Develops systems capable of and application in
predicting outcomes and specific contexts by
learning patterns solely from human
input data. decision-makers.
● Functions with the automation ● Requires careful
of decision-making within the consideration and
SCRM. validation of insights
● Assists in identifying and provided.
understanding various risks ● Potential biases in
within the SC. the data used for
● Recognizes and extracts training ML.
patterns related to risks. ● Complexity in
● Enables assessment of levels of implementing and
risk present. maintaining ML
● Can be used in trend forecasting systems.
and inventory accuracy. ● Potential for errors
● Analyses extensive amounts and inaccuracies in
and data. predictions.
● Increases efficiency in
decision-making.
● Increases resilience in the SC.
Cloud computing
● Provision of reconfigurable ● Complexities and
resources such as software, challenges,
infrastructure, or platforms. particularly
● Shared pool of virtualized concerning
resources. cyber-related
● On-demand self-service, broad disruptions.
network access. ● Heightens risks
● Enhances structural flexibility. associated with cyber
● Improves responsiveness. threats.
● Competitive advantages. ● Requires systematic
● Facilitates prevention and analysis of potential
35
management of disruptions. risks for improved
● Enhances operational resilience defenses.
and SC collaboration. ● Highly dynamic
● Cost reduction environment,
● Scalability maximization and contributing to
rapid development. operational risks.
● Enhances flexibility, agility and
efficiency within the SC.
● Enables coordination
mechanisms such as protection,
response, and adaptation.
Internet of Things
(IoT) ● Facilitating interactions by ● Data security and
linking numerous devices. privacy concerns.
● Enables information exchange ● Interoperability
among various smart devices, issues due to the
users and data centers. absence of universal
● Discerns relationships between standards.
entities and recognizes complex ● Integration
patterns. challenges with
● Enhances efficiency and legacy systems in
reduces costs. many supply chains.
● Ensures product quality and ● Resource-intensive.
promotes sustainability in the ● Complex and
SC. expensive
● Enables real-time tracking and implementation and
data driven decision-making. integration.
● Aids in detection and resolution ● Requires the right
of inconsistencies and safety expertise and
crises. competencies.
● Reduces uncertainties and ● Ensuring safety and
facilitates demand-supply information security
synchronization. is challenging.
● Allows organizations to adapt ● Requires substantial
quickly to changes in demand bandwidth and cloud
and optimize inventory levels. processing.
● Improve demand predictions,
accuracy and forecasting.
● Facilitates real-time decision
making.
Artificial Neural
Network (ANN) ● Computes anticipated costs and ● Challenges with
production loss risks. imbalanced data,
● Discerns correlations between incomplete data,
input and output.
36
● Eases supplier selection high dimensionality
process by evaluating reliability and limited
and providing flexibility. accessible data for
● Identifies, comprehends, and learning.
predicts patterns from vast ● Complexity and
datasets. interpretability
● Widely applicable within the difficulties withlarger data sets.
SC. ● Suboptimal results
● Enhancing accuracy and with an excessive
performance in the SC. number of hidden
● Enables ability to identify and nodes in the ANN
proactively address risks and model.
disruptions in SCM.
● Handles incomplete data.
● Mitigates uncertainties.
● Recognizes patterns and predict
potential disruptions with high
accuracy.
● Provides deeper insights and
more accurate depictions of
underlying connections and
patterns.
● Outperforms regression
methods in predicting
completion times.
● Enhances SCM performance.
● Indicate variables influences.
Agent-based
system ● Replicated behaviors and ● Complexity in
interactions of autonomous modeling behaviors
agents. and interactions of
● Adapts behavior to anticipate, autonomous agents.
respond, enhance flexibility and ● Challenges in
recover from disruptions. accurately
● Provides proactive responses to representing
mitigate disruptions in real-world SC
real-time. dynamics.
● Improves coordination. ● Resource-intensive
● Optimizes time and resource in terms of
usage. computational power
● Evaluates the impact of and data
disruptions and devises requirements.
strategies for managing ● Requires careful
disruptions. validation to ensure
37
● Streamlining decision-making. simulation accuracy.
● Facilitates mitigating and ● Limited ability to
managing uncertainty and capture the full
conflicting information during complexity of human
disasters. decision-making and
behaviors.
● Potential for
unexpected emergent
behaviors in the
simulation.
● Interpretability
challenges in
analyzing the results.
Blockchain ● Enhances trust among SC ● Unclear regulation
actors regarding data exchange. and laws.
● Ensures confidentiality and ● High operation and
openness for data access and implementing costs.
updates. ● Interoperability
● Improve traceability of challenges leading to
information, processes, and standardization
financial flows. issues.
● Increases visibility in the SC. ● Unclear regulations
● Reduces administration costs hindering innovation
and issues regarding fraud. and development.
● Requires change
management within
the SC.
● Conflicting
objectives among
stakeholders.
● Technical issues
involving scalability,
security, power
consumption and
privacy.
38
2.3. Supply Chain Risks
2.3.1. Regulatory and Compliance Risks
In recent years, supply chain compliance has evolved to become an integral component of
traditional compliance networks due to its expansion to include various concepts and
principles. The attributed reason for the expansion can be the shift from shareholder
capitalism to stakeholder capitalism. Supply chain compliance can be explained as
encompassing adherence to legal standards at every stage of the supply chain, that includes
all the suppliers and all processes involved in manufacturing products and delivering
services. Effective risk mitigation requires close coordination among all departments
involved, such as legal and compliance, procurement and sustainability departments. With
these developments in mind, conducting thorough checks on supply chain processes will be
seen as a pivotal activity for organizations in order to uphold compliance and standards
(Ersoy & Akdag, 2021) Accurate communication among supply chain elements is essential
for optimizing supply chain management operations. This leads to improved performance,
agility, and adaptability, while also reducing uncertainties among supply chain partners that
can be caused due to sharing barriers. Supplier self-assessment, conducted through informal
audits, serves as a method for supply chain organizations to enhance transparency and gain
trust (Afrifah et al., 2022).
It is crucial for organizations to have a comprehensive understanding of relevant regulations
and ensure that their processes are designed to align with them. Regulations can be intricate
in terms of their conditions, targets, and scope. Ensuring regulatory compliance is therefore
not only labor-intensive but also time-consuming and complex. While business processes are
crafted to guide and regulate operations, governmental regulations are formulated and
enforced to uphold social welfare, protect local interest, and regulate business activities
(Jiang et al., 2015). There are three types of regulatory policies: command-and-control
regulation, market-based policies, and non-regulatory approaches. Command-and-control
regulations, also referred to as traditional regulations, use coercive measures such as bans on
production or the use of certain substances. Market-based policies utilize market mechanisms
and prices to incentivize companies to reduce negative environmental impacts, often through
taxes or charges. Non-regulatory approaches aim to encourage companies to reduce their
environmental impact voluntarily through programs that provide information without
imposing mandates (Darnall et al., 2019).
One primary goal is to ensure compliance across various dimensions, including production,
environmental and labor standards throughout multi-tier supply chains. Firms are extending
their sustainability initiatives to lower-tier suppliers by leveraging their leadership practices
and supply chain knowledge. However, conflict may arise due to differing social, ethical and
environmental standards among countries, leading to efficiencies in the supply chain.
Addressing these challenges requires coordination among supply chain actors to harmonize
standards, but this is complicated by non-technical factors such as politics, trade wars, and
conflicts (Wang et al., 2023).
39
Many organizations have enhanced their corporate social responsibility efforts, particularly in
environmental and social sustainability. These efforts involve improved visibility and
transparency within the supply chain to demonstrate commitment to the public and
customers. However, disclosing supply chain information poses risks, including the potential
exposure of competitive advantages or vulnerabilities, association with unfavorable supplier
practices, and suspicion from consumers or advocacy groups about hidden information
(Sodhi et al., 2019). Kulkarni et al. (2021) proposes an AI-aided model-driven approach to
regulatory compliance, which shifts the focus from document to models, reducing analysis
and synthesis processes. This approach enables the authoring, validation, and compliance
checking of regulation models, allowing for simulation through real-time scenarios to
quantify risks. Artificial intelligence is also becoming increasingly applicable in corporate
compliance, offering predictive capabilities through big data analysis and real-time risk
alerts. AI can aid in identifying and preventing misconducts, improving monitoring, reporting
and enhancing corporate risk assessment by detecting risk areas and predicting future events
(Sabia, 2019).
2.3.2 Supplier Risks
Suppliers constitute a critical role in determining the success of an organization, particularly
given the ongoing globalization and the trend of outsourcing and contract manufacturing.
Their potential faults or sudden disruptions can cause a negative impact on the performance
of their customer’s businesses. Supplier risk is defined as “an unexpected event originating
from an upstream supplier that spreads its effects downstream in the supply chain” (Sarker et
al., 2016). These risks can be categorized into various types: economic risks stem from a
supplier’s lack of financial stability or insolvency, environmental risks encompass
uncertainties within the supply chain environment including accidents and natural disasters,
and cost-related risks involve changes in the prices of purchased goods due to scarcity or
market unavailability. Other risks pertain to quality failure, where delivered goods from the
suppliers are unable to meet the firm’s quality specifications and requirements. Many risks
tend to appear in clusters, with some risks causing or exacerbating others. To mitigate these
risks, a vital aspect of risk management is the identification of risks, which involves the
firm’s ability to understand the dynamics of the supply chain and predict potential
disruptions. This enables the firm to assess knowledge about the nature, location and
conditions of the disruptions and risks. Increasing visibility among affected suppliers within
the supply chain aids to reduce uncertainty by providing more information about the risks,
thus enhancing understanding of the nature, location and conditions of the risks and
disruptions (Sarker et al., 2016).
Perceptions of risks vary among suppliers because organizations operate under their unique
conditions and act rationally within the bounds of the information available to them. Risks
are often managed in silos, with different departments of the organization addressing risks
within their respective domains. However, differences in the visibility levels of various
suppliers impact how different actors within an organization manage or respond to risks. To
40
effectively manage these silos, organizations must understand the dependencies among risks.
Positive dependencies suggest that managing risks in silos can be efficient, while negative
dependencies indicate potential counterproductivity. Therefore, cross-functional integration
and communication are proposed as measures to increase internal risk visibility among
stakeholders. Internal visibility refers to the visibility of supplier risks among various actors
(Sarker et al., 2016).
There are four influences that affect supplier risk, performance and disruption within a supply
chain. These influences encompass supplier attributes, which include financial performance,
human resource factors, cultural factors, and relationships. Additionally, supply chain
strategy and structures constitutes a role, encompassing supply chain type (lean, agile or
hybrid), supplier type, business structure and geographic location. The last two influences
relate to endogenous uncertainty, including market turbulence involving factors such as price
sensitivity, introduction of new products and demand fluctuations and technology turbulence
referring to rapid changes, interest rates, inflation, strikes and disasters. Determining supply
chain risks and managing their outcomes can be enhanced through the relationship between
the focal firm and its suppliers. The characteristics of these relationships should incorporate
formalization, dependency, trust, information exchange and financial performance (Lavastre
et al., 2014).
Collaboration among suppliers within the supply chain enhances the risk management
capabilities of firms. It facilitates and improves communication between firms and tier
suppliers, allowing for more efficient management of changes in daily operations and
enhancing the ability to mitigate uncertainties in the supply chain. Through collaboration
firms are able to better avoid, reduce and mitigate potential risks and disruptions. For
instance, organizing meetings with key suppliers to discuss the latest purchasing plans
enables suppliers to better reach the requirements and standards of the focal firm regarding
product quality, quantity and price. This results in reduced internal processing of risks within
the firm. The level of supplier collaboration varies depending on the resource commitment
and the intentions of the partnerships (Shuting & Chen, 2019).
A comprehensive approach to supplier risk management has been developed, comprising five
main steps. Similar to conventional risk management practices, the primary focus of
mitigating risks and disruptions lies in identifying various supplier risks. Risks associated
with critical and strategic sourcing are prioritized for the risk management process, as
disruptions to these flows can significantly impact the firm’s market position. The second
step involves assessing supplier risks employing a two-sided perspective rating mechanism.
This mechanism incorporates both internal firm rating and external supplier rating, which are
combined to represent the supplier risk structure. The strength of this approach lies in its
ability to consider the viewpoints from both the perspective of the firm and the supplier,
allowing for the presentation of differing opinions and creating a comprehensive overview of
the risks associated with a particular supplier. Additionally, the third step of this approach
involves reporting and decision-making regarding detected supplier risks. This entails
aggregating and classifying supplier risk data, enabling the categorization of suppliers into
41
groups associated with high or low risks. It is imperative that the firm evaluates responses
from the organization addressing the calculated supplier risk results. These responses are
intended to improve the risk performance of the supplier bases and may include initiatives
such as information sharing, performance standards, and partnership programs. Such
management responses are considered enablers of risk mitigation, supporting trust building
and establishing a collaborative relationship among supply chain partners. The final stage
focuses on supplier performance outcomes. The objective of this stage is to reduce the
inherent risks associated with suppliers and enable them to meet the manufacturing
company’s supply needs (Matook et al., 2009).
Supplier selection constitutes a crucial role within supply chain management, influencing an
organization’s resilience and risk management strategies. To ensure operational continuity
and mitigating disruptions, decisions regarding which suppliers to select are essential (Setak
et al., 2012). Supplier selection aims to ensure that suppliers reach the requirements and
standards of the focal firm, which is achieved through the assessment of supplier sites,
processes, and compliance with criteria of the selection (Leisa et al., 2018). Supplier selection
is considered foremost for achieving operational excellence and enhancing supply chain
resilience. A well-structured supplier selection process considers various critical criteria
including digital collaboration, risk mitigation strategies, supply chain integration
capabilities, quality, compliance assurance and cost optimization processes (Sahoo et al.,
2024). Common criteria encompassing quality, lead time and price are prioritized, other
factors such as the capacity of production facilities, delivery, labor practices, environmental
compliance and sustainability actions have become equally important (Ho et al., 2011).
Failure to consider all relevant selection criteria may pose significant risks to the firm’s
supply chain. Therefore, structured risk assessment approaches should be applied consistently
throughout the selection process to mitigate and reduce supplier-related risks (Leisa et al.,
2018). Selecting the most appropriate supplier not only reduces purchasing costs but also
enhances corporate competitiveness and overall supply chain performance (Ho et al., 2011).
To assess supplier performance objectively and subjectively, a combination of mathematical,
statistical and AI methods, such as Explainable Artificial Intelligence (XAI), can be
employed. XAI enhances transparency, interpretability, and trust in decision-making,
enabling business to align choices with strategic objectives and continually enhance the
evaluation process (Kumar & Kumar, 2023). Machine learning techniques, including decision
tree and Artificial Neural Networks (ANN), are particularly effective in handling complexity
and uncertainty within supplier selection. These methods leverage historical data, current
conditions, and future projections to predict supplier performance accurately and refine
decision-making over time (Atwani et al., 2020). Cognitive sourcing platforms like Silex
streamline the procurement process by facilitating supplier selection from a vast pool of
options. By integrating data from suppliers and external sources, such platforms provide
comprehensive insights into industrial, financial, and technological aspects, empowering
customers to make informed decisions and access diverse information when selecting a
supplier (Allal-Chérif et al., 2021).
42
2.3.2.1. Monitoring of Suppliers
Monitoring, measuring and evaluating supplier performance is considered a time-consuming
task. However, technological innovations such as RFID, Internet of Things and Big Data
facilitate data assessment, making this process become more efficient. The quality of the
relationship between the buyers and suppliers is positively correlated with the monitoring of
the suppliers, potentially enhancing the buyer’s performance (Maestrini et al., 2018). By
carefully selecting suitable criteria for supplier selection and consistently monitoring supplier
performance, purchasing departments can effectively translate their involvement in strategic
planning into improved purchasing performance across various dimensions including cost,
quality, delivery, flexibility and innovation. The supplier performance evaluations are often
based upon criteria including cost, delivery, and flexibility performance, which is related to
the supplier’s ability to adapt and adjust to sudden changes in quantity requirement, deliver
product on short notice and produce smaller production at higher frequency (Nair et al.,
2015).
Supplier portfolios have become increasingly complex and challenging to manage due to
variations in size, location, organizational structures, and corporate cultures. Purchasers can
analyze vast amounts of data to monitor supplier performance and assist them in improving
based on specific criteria. Additionally, purchasers must ensure that strategic suppliers are
satisfied with the relationship through regulatory discussions about the quality of interactions
(Maestrini et al., 2018). Some manufacturing firms utilize AI to be aware of their rating by
suppliers through software applications employing algorithms to calculate ranking. Being one
of a supplier’s best clients in terms of relationship quality could provide certain privileges
such as priority in the event of emergencies and disruptions. Moreover, AI enables a more
agile, reactive, and efficient approach to purchasing, allowing buyers to focus on strategic
missions and source products efficiently. AI can further provide recommendation measures to
benefit from opportunities and mitigate potential threats by comparing external and internal
data, revealing inconsistencies in supplier’s practices (Allal-Chérif et al., 2021).
Several global companies utilize advanced AI technologies for planning and adapting to
supply chain disruptions. These tools enhance visibility into the supply chain, facilitate faster
responsiveness, and deepen relationships with suppliers by expanding purchases to new
items. For instance, Unilever employs an AI application called Scoutbee to find alternative
sources of supply at short notice, providing potential new suppliers by analyzing factors such
as their finances, customer rating, sustainability scores, and customs documents. Another
approach is to adopt AI to assess whether existing suppliers can provide additional materials.
Koch Industries, one of America’s largest privately held conglomerates, leverages on an AI
tool developed by Arkestro, to optimize its supplier base, generating supply options, often
among existing suppliers (Van Hoek et al., 2023).
The implementation of artificial intelligence within supplier relationship management fosters
real-time communication, provides predictive insights into supplier performance, and
improves collaboration among the actors. Predictive analytics enable improved forecasting of
43
demand, allowing suppliers to align production and distribution with the organization’s actual
needs, minimizing disruptions and optimizing inventory levels. However, companies must
mitigate and manage various challenges for successful implementation of such technologies,
including organizational resistance to change, data security concerns, and the need for a
skilled workforce (Glory, 2023).
2.3.2.2. Prediction and Identification of Supplier Risks
Digital supply chain surveillance (DSCS) is the proactive monitoring and analysis of digital
data, enabling organizations to extract supply chain-related information without explicit
consent from involved parties. Artificial intelligence has enhanced and streamlined DSCS,
allowing it to scale up and present significant opportunities in automating the detection of
actors and dependencies within the supply chain. This aids firms in identifying risky,
unethical, and environmentally unsustainable practices. Additionally, DSCS supports various
areas including visibility, sustainability, and resilience by employing machine learning
approaches and predictive algorithms. It encompasses three phases: data collection and
processing, data analysis, and extraction of actionable insights. The first phase involves
selecting appropriate data sources and processing data, while the second phase requires
applying suitable algorithms to derive relevant statistical patterns. The third phase involves
extracting applicable conclusions and information from the data to enhance decision-making
(Brintrup et al., 2023). DSCS proves considerably useful, enabling companies to assess
information and data independently and thus simplifying risk monitoring complexities
(Brockmann et al., 2022).
Several DSCS approaches have been proposed to predict supply chain risks, such as
analyzing data from social media platforms such as Twitter, to monitor supply disruptions.
Furthermore, other applications include employing classification algorithms to predict
supplier delays using historical delivery data, which can be adopted to optimize inventory and
safety stock levels. The implementation of DSCS within supply chain management requires
careful consideration of uncertainties in predictions against the costs of acquiring such
information and defining appropriate actions. DSCS does not dictate actions but rather
provides evidence to support decision-making (Brintrup et al., 2023). One commonly adopted
approach within DSCS is extracting supply chain network information from publicly
available data. Machine learning can facilitate this assessment by automatically inferring
relationships between entities. Another approach involves utilizing machine learning to
predict relationships and connections within incomplete knowledge in the supply chain
network (Brockmann et al., 2022).
Apart from DSCS, another method for mitigating disruptions is data analytics, which can
predict first-tier supply chain disruptions using historical data. Data analytics encompasses
descriptive, predictive and prescriptive analytics. Descriptive analytics pertains to current
supply chain metrics such as total inventory stock, annual sales fluctuations and average
customer spending. Predictive analytics utilize algorithms to forecast future business states,
such as customer purchasing patterns and sales trends. Prescriptive analytics build on
44
prediction to optimize current states and take appropriate actions toward more desirable
outcomes, through predicting trade goods’ volatility, adjusting prices or inventory, avoiding
predicted traffic disruptions and adjusting warehouse inventory levels. Supply chain analytics
can be applied within procurement to mitigate and manage supply risks and supplier
performance, enabling proactive responses to supply chain risks (Brintrup et al., 2019).
Baryannis et al. (2019) propose a framework for enhancing supply risk prediction by
integrating artificial intelligence into the risk management process. The framework
emphasizes interaction and synergy between the supply chain actors and AI experts, where
decisions by the experts depend on specific input from the supply chain, and any
models/results produced need to be interpretable to influence risk management
decision-making. Successful implementation relies on meaningful communication between
the two parties. This framework can predict supply chain disruptions utilizing external and
internal data to quantify causal effects on delivery reliability. It contains the ability to learn
casualties identified within observed data and analyze interventions affecting supply chain
disruptions. Identifying causal relationships offers potential to identify deficient supplier
relationships and bundle supply chain risk management activities. Furthermore, it estimates
associations between supplier characteristics and supply disruption frequency, aiding
managers in prioritizing and executing strategic decisions to optimize their supply chains.
2.3.3. Resilience Planning and Strategy
AI implementations enable predictable methodologies that mitigate potential risks or the
occurrence of disruptions in the supply chain, which makes it possible to differentiate certain
patterns of data that can be used for further analysis to achieve a deeper insight into the
operational processes and the identification of necessary improvements. Therefore, AI in risk
management and resilience in the supply chain is an ideal integration area where this analysis
can likewise be used to help predict the effects of external events on an organization (Riahi et
al., 2021).
One key approach in order to achieve a successful and complete supply chain risk
management is to invest and develop technologies to enhance visibility and resilience through
the supply chain. Artificial intelligence and advanced analytics are some of the technologies
that can ease decision-making and enhance supply chain management. These technologies
enable experts in supply chain management to gain insights into the practical data and with
that gain competitive advantage. One argued reason behind the adoption of risk management
in organizations is primarily due to the intensified global competition and advancements in
technologies. Thereby, supply chain risk management also allows for new technologies to
manage risks and improve performance (Emrouznejad et al., 2023).
Supply chain risk management with AI involves conducting scenario-based assessments,
generating disruption models encompassing strikes, natural disasters, and pandemics. By
continuous and detailed analysis of these risks, businesses can develop robust strategies to
maintain operational continuity (Richey et al., 2023). According to Baryannis et al. (2019)
there are two forms of implementation strategies of supply chain risk management, one which
45
follows a reactive strategy and the other a proactive strategy. The distinction lies in the
reactive strategy’s capability to evaluate the implications after the occurrence, whereas the
proactive strategy is to enhance any preparation before the disruption and mitigate the impact
on the supply chain. The proactive strategy in combination with AI techniques enables
organizations to anticipate risks and their implications with a higher accuracy (Baryannis et
al., 2019).
Resilience within the supply chain refers to its ability and capacity to effectively manage
disruptive events and promptly restore its previous performance levels. Supply chain
resilience empowers the supply chain to ensure an uninterrupted supply of products and
services to customers in a turbulent and uncertain environment. The utilization of artificial
intelligence can improve the dynamic capabilities of sensing, seizing, and transforming,
thereby mitigating the risks of actual disruptions and proactively preventing potential future
issues from appearing (Belhadi & Kamble et al, 2021). AI constitutes a crucial tool in
enhancing resilience within the supply chain through continuous monitoring and minimizing
the impact of unforeseen disruptions (Ganesh & Kalpana, 2022). It also contributes to
organizational performance and resilience by facilitating real-time connectivity among
stakeholders (Modgil et al., 2022). In supply chain risk mitigation, AI contributes to
readiness, response, recovery and growth. The supply chain faces risks of various
consequences derived from a lack of resilience such as decreases in revenues, loss of sales,
damage to brand reputation, lowered customer service and increased supply chain and
shortage costs (Kassa et al., 2023). Additionally, a lack of resilience within the supply chain
can result in a decline in the firm’s performance, even with extensive AI usage (Sullivan &
Fosso Wamba, 2022).
2.4. Implementation Strategies for AI
It is considered crucial for businesses contemplating the implementation of new technologies
within their organization to initiate communication with employees early on, informing them
about the impactful changes and aiding in their successful integration. Additionally, it is
recommended to establish a robust return on investment (ROI) calculation and implement
cost monitoring before investing in and adopting new technology to ensure that the
transformation of the business is successful (Hangl et al., 2022). Two main factors contribute
to the successful implementation of AI in an organization. First, there must be firm
commitment to fostering a culture of data-driven decision making. Second, leadership needs
to consistently invest in AI training and competencies across the entire company (Shirvastav,
2022). The influence of the organizational antecedents such as top-management commitment,
involvement of employees and stakeholders and experience of automation projects will
impact the success of automation (Nitsche et al., 2021). Most companies have difficulties
adopting AI successfully due to ineffective change management and vague understanding of
industry 4.0 and its strategic importance, emphasizing the requirement of collaboration and
integration of stakeholders and establishing a high maturity level of the desired technology to
implement (Cannas et al., 2023).
46
The primary objective of implementing AI in businesses is consistent: to increase revenues,
enhance efficiency, and develop new data-enabled offerings. However, many companies face
uncertainty and confusion about where to integrate AI and how to harness its advantages. The
recommended starting point involves optimizing current business processes using AI,
leveraging existing internal data sources to improve business models, internal processes,
products, services and functions. Success in implementing AI relies on management and
organizational understanding of data and AI applications, with a strategy contingent on
awareness and incorporation of the organization’s business goals. The priorities of AI should
align with business priorities, focusing on facilitating informed decision-making, faster
information retrieval, and process automation, rather than solving broader business model
issues. A committed leadership is considered a key determinant of achieving success in
digital transformation and the transition to a data-driven company (Kruhse-Lehtonen &
Hofmann, 2020).
Once the firm assesses knowledge of its goals and AI uses, exploring new data-driven
business opportunities is crucial and suggested. This includes treating data as a business
commodity (selling data) and forming data partnerships (creating offerings by pooling data
from various organizations). High-quality data, structured according to FAIR
(Findable-Accessible-Interoperable-Reusable) principle, is essential for successful,
productized AI. However, an organizational disconnect often exists between IT teams, data
engineers and business functions using data-driven insights. Conducting data due diligence is
necessary to understand the current state of data assets, addressing questions about data
existence, location and accessibility (Kruhse-Lehtonen & Hofmann, 2020).
A different approach suggests that the business should initiate the implementation of the AI
process with data assessment and preparation. This involves two primary activities: collecting
data from industrial sensors and entire IoT systems, and organizing and structuring the
gathered data. This process includes eliminating duplicates, irrelevant records, handling
missing data attributes, and labeling the data necessary for the AI’s learning process.
Furthermore, the second step of this strategy involves focusing on the learning and training
aspect of the AI. Artificial intelligence can acquire knowledge through repeated learning,
resulting in for example acceptable forecasting accuracy, irrespective of the data set’s size.
The training data set must encompass all the features required for the model, aiming to
construct an accurate relationship based on available data and current algorithms. Lastly, the
model should be deployed within the actual business process in a real-life environment. The
testing data set is then utilized to validate the accuracy of the generated AI model. The model
must undergo revalidation in the latest environment and be optimized based on the given
results and feedback (Helo & Hao, 2022).
Adopting artificial intelligence in an organization demands substantial organizational
changes, affecting the configuration of supply chains, decision-making procedures, and the
overall structure. Navigating this dynamic environment requires organizations to find a
delicate balance between sustainability initiatives and digital advancements, resulting in
alterations to their business models. However, innovating business models comes with
47
challenges, involving resource demands, capital requirements, and inherent risks. The
practical application of an AI system presents difficulties, relying extensively on the expertise
and capabilities of logistics and supply chain management professionals for a thorough
feasibility evaluation (Richey et al., 2023). This digital transformation necessitates the
training and ongoing professional development of employees, coupled with the adoption of
lean production methods and continuous improvement approaches in production processes.
Enhancing and refining infrastructure and fostering collaboration among supply chain
partners become crucial for promoting openness and adaptability, thereby facilitating the
successful implementation of AI in the organization (Cannas et al., 2023).
2.5. Gaps in the literature
Several gaps and areas for future research within the literature have been identified.
Emrouznejad & Sıcakyüz (2023) highlight the need for a broader industry perspective
research beyond the internal operations of the supply chain, emphasizing the necessity for
additional studies on risk factors and new insights into risk management. Baryannis et al.
(2019) further underscore the potential for future research within the field of supply chain
risk management and AI, while Hudnurkar et al. (2017) stress the importance of exploring
supply chain risk classification. Moreover, Guida & Cantano (2023) proposes delving deeper
into supply chain risk management, particularly regarding the influence of AI, focusing on
the nature of data and AI technologies that facilitate dynamic resource reorganization to
address supply chain risks. Bode & Wagner (2014) advocate future empirical research on
identifying supply chain characteristics that increase the frequency of disruptions.
Additionally, Jahani et al. (2021) highlights the need for a comprehensive framework on how
Industry 4.0 technologies create value for procurement departments, while Giannakis et al.
(2019) propose investigating cloud-based approaches and cloud-level interoperability.
Obstacles to AI adoption in procurement departments, such as competencies, culture, and
maturity, should be further analyzed (Guida et al. 2023). Blos et al. (2015) suggest studying
the ability of agent-based supply chains to mitigate disruptions, while Richey et al. (2023)
identify a gap in understanding the implications of AI implementation on the legal framework
and regulatory requirements of supply chains, particularly for international operations.
In contrast, Toorajipour et al. (2021) argue for industry-specific studies rather than
generalized applications of AI techniques. Li et al. (2021) advocate larger sample sizes and a
broader variety of network structures to deepen the understanding of disruption propagation,
while Sahoo et al. (2024) emphasize the need for practical insights into the application of
Industry 4.0 technologies within the area of supplier selection. Sharma et al. (2022) highlights
the underutilization of AI in addressing complex supply chain problems, emphasizing the
focus on well-defined areas like network design and forecasting. Hangl et al. (2022)
underscores the need for detailed research on the robustness and interpretability of AI
utilization in the supply chain, particularly exploring combinations with other Industry 4.0
technologies for improved planning and optimization.
48
The identified gaps from the literature align with the gaps identified in this research. There is
limited literature that delves into and focuses on risk factors and insights into risk
management, highlighting the need to broaden the understanding of AI’s influence within
supply chain risk management. Furthermore, other identified gaps include identifying
characteristics that increase disruption frequency within the supply chain, analyzing potential
obstacles to AI adoption in procurement and investigating AI’s role in addressing complex
supply chain problems. Additionally, there is limited literature exploring AI’s ability to
predict and identify risks and disruptions within suppliers, with most research focusing on
AI’s role in supplier risk management. None of the assessed literature in this research focused
on a specific company within a particular industry; instead it generally addressed industries
such as healthcare or e-commerce. This research addresses this gap by presenting a
comprehensive framework of approaches and strategies for utilizing AI within the supply
chain to mitigate disruptions and risks, applied to a global company.
3. Methodology
This chapter will provide the necessary and sufficient understanding of the method that has
been applied throughout this research i.e., research design, data collection and Validity and
Reliability, which provides a concise overview for the reader of the methodological approach
applied.
3.1. Research Design
A qualitative research approach was adopted for this study. Due to this study aiming to
contribute with valuable insights and to facilitate the development of more resilient supply
chain systems, this approach was most suitable due to its ability to allow the researcher to
explore and provide deeper insights into real-world problems (Tenny et al., 2017). This
qualitative research study involves a comprehensive literature review to contextualize the
study within the application areas of AI in supply chain risk management and thus how to
enhance resilience. Following this, an empirical analysis will be conducted both on primary
sources and secondary sources, allowing for interviews and previous studies published within
this research topic. Interviews were considered more appropriate to conduct due to it allowing
for depth in the analysis in comparison with written questionnaires (Bengtsson, 2016).
Limitations inherent in the study will be acknowledged, alongside measures to enhance
reliability and validity. The research aims to answer the research questions applied which will
be pursued through a methodological approach designed to ensure the accuracy and integrity
of the findings.
Ensuring ethical standards remains paramount throughout this research process. Key
measures include obtaining participants’ consent, safeguarding personal information with
confidentiality, and maintaining ongoing dialogue and communication throughout the
research journey. Ensuring ethical considerations in research involves four essential
requirements. Firstly, researchers must provide clear information about the study’s purpose to
49
those affected. Secondly, participants must give their informed consent. Thirdly, all personal
data must be handled with utmost confidentiality. Furthermore, information about participants
should be used solely for research purposes (Patel & Davidsson, 2019). Furthermore,
voluntary participation has been of great importance as an ethical principle in this research,
ensuring the free will of the individuals whether to take part without coercion. Adequate
information has been provided before and upholded during the study, to ensure that these
principles respect the participants’ autonomy and maintain the integrity of the research
process. As per Collis & Hussey (2021), voluntary participation stands as an paramount
ethical principle in human-involved research, particularly within academic domains. It is
further underscored the significance of refraining from providing any financial incentives in
order to participate as this potentially leads to biased findings.
3.2. Data Collection
To assess and gather data for this research, a combination of literature review and empirical
data collection through interviews was conducted. The initial phase involved an extensive
literature review where relevant academic papers and conference documents were searched
through databases such as GoogleScholar, SuperSök and ScienceDirect. Researchers were
conducted using relevant keywords and search terms to ensure that all relevant research on
the topic was covered. For instance, such keywords encompassed supply chain, disruptions,
AI technologies, forecasting, risk mitigation, risk management, resilience, supply risks, AI for
risk mitigation. Furthermore, the results were filtered based on relevance and publication to
ensure only current and relevant research was included.
The empirical data assessments were conducted utilizing semi-structured qualitative
interviews, facilitated through the digital platform Microsoft Teams. The questions were
formulated and sent out to the respondents before the date of the interview, allowing the
participants to prepare and provide nuanced responses. Given the research significance of
providing valuable insights and enhancing the resilience of supply chain systems, this
interview approach deepens the analysis of quantitative data and provides a nuanced
understanding of participants’ perspectives (Adeoye-Olatunde & Olenik, 2021). Another
advantage of employing this qualitative interview method is the flexibility it affords the
researcher to tailor follow-up questions based on respondents’ responses and create space for
participants to elaborate. Pre-determined and formulated questions provided a structured
framework for the interviews, ensuring focused discussions aligned with the research
objectives and preventing the collection of extraneous data (Kallio et al., 2016).
Finally, data from the literature review and empirical sources were integrated and synthesized
to develop a comprehensive understanding of the research topic. Results from different data
sources were triangulated to validate and confirm key insights and conclusions, enhancing the
credibility and robustness of the research findings. By applying a comprehensive method of
data collection that combines literature review and empirical data collection through
interviews, this research aims to generate robust and valuable insights within the research
area, contributing to knowledge advancements in the field.
50
3.2.1. Validity and Reliability
Validity within qualitative research encompasses the entire research process (Patel &
Davidsson, 2019). It indicates the likelihood that the study findings are accurate and not
influenced by biases (Khorsan & Crawford, 2014). This is determined by the interpretation
of the data obtained as a result of the analysis (Sürücü et al., 2020). Two types of validity
have been identified: internal validity, which assesses the sustainability of the relationship
between targeted variables, and external validity, which concerns the generalizability of
findings (Bryman & Bell, 2011). Overall, it is about the researchers’ ability to assess
qualitative data to produce a comprehensive and reliable analysis and interpretation of the
study’s subject matter (Patel & Davidsson, 2019).
Reliability refers to the research’s ability to yield similar results upon replication (Sürücü et
al., 2020), focusing on consistency, accuracy, equivalence, and homogeneity
(LoBiondo-Wood & Haber, 2014). Two categories of reliability have been identified: internal
reliability and external reliability. Internal reliability pertains to the consensus among the
research team regarding the interpretation and understanding of assessed information and
data. On the other hand, external reliability concerns the extent to which the study’s findings
can be generalized to other social environments and situations (Bryman & Bell, 2011).
Internal validity in the research has been solidified through consensus and agreement on the
interpretation of assessed data, coupled with ongoing dialogue and communication
throughout the research process. Daily meetings and regular contact have eliminated
confusion and misunderstandings regarding interpretation and analyzation of data. However,
the generalizability may be constrained by the focus on Volvo AB’s supply chain and the
specific adoption and application of AI within their unique supply chain. Conversely, the
diverse technologies and AI application areas within various risk management domains may
be transferable to other industries or companies.
Given that validity encompasses the entire research process (Patel & Davidsson, 2019), it has
been addressed at every stage. To ensure validity, triangulation has been adopted.
Triangulation involves employing diverse data assessment methods that are subsequently
interconnected to offer a thorough comprehension of the study’s subject matter (Patel &
Davidsson, 2019). It is employed and embraced to mitigate the impact of potential researcher
bias and to ensure confirmability, where procedures are clearly defined for verifying and
cross-checking data throughout the study (Renz et al., 2018). In this study, data assessment
methods are approached from both qualitative perspectives, through semi-structured
interviews, and theoretical perspectives, where multiple sources shed light on and analyze the
same topic from different viewpoints. This triangulation fosters a robust understanding of the
subject matter, thereby enhancing the research’s validity.
4. Empirical Findings
This chapter presents the answers conducted from the interview with Volvo AB.
51
How are AI technologies today being applied in Volvo AB’s supply chain?
AI technologies are not used extensively within Volvo AB’s supply chain. We try to
incorporate them in our scope of risk management. It is employed in the sourcing, automating
manual work and for analyzing code.
How do you mitigate disruptions and risks today?
To mitigate disruptions and risks we audit suppliers, set up certain requirements for them to
fulfill and need ISO certifications. Furthermore, we ensure contact with our suppliers and
shipping companies on a regular basis. We continuously try to find new sources to retrieve
information and market changes are detected by our AI solution, Prewave. Furthermore, we
analyze our risks.
How do you analyze risks?
Prewave discovers external risks from capturing data. We also have a risk dashboard where
particular scoring and impact points are visualized. Depending on what the risk is about,
different persons are working with it. We want to discover how AI perhaps can make this
easier.
What are your expectations of utilizing AI within supply chain management?
As of right now, we have a lot of external data that are being collected. We would like to
implement AI to filter the noise and sort out the relevant information. We could manage to do
it ourselves, but it would require a lot of resources.
When implementing AI did you need to alter the supply chain or did the AI fit into the
supply chain configuration directly?
The implementation requires that each supplier needs a certification, which is not possible
due to us having 15000 suppliers. We need an AI that is suitable for our supply chain, rather
than our supply chain being suitable for a specific AI.
What challenges do you see in implementing AI in supply chain management?
The challenges we see regards data quality, availability of the right data. This is something
we are struggling with today. We need the right applications.
Have you considered any specific AI technology or tool to implement as of today that
would fit your supply chain? For example the internet of things?
We have internet of things in our trucks for maintenance purposes. We believe that AI is more
challenging to implement. Blockchain could be an alternative, but it seems to be a challenge
due to us having a complex supply chain.
Would you like to implement blockchain? Or is it too challenging?
52
At the moment we are depending on the contract of suppliers. If not in contract, we cannot
force them to implement blockchain and therefore we are currently using Prewave to make
predictive analysis. We are thinking about integrating AI to improve visibility.
What do you believe is the future for AI in the supply chain? New unexplored
application areas?
The supply chain visibility part is the key and a very interesting field to us. AI and blockchain
can give a great overview. We believe it could improve aspects of sustainability and allow us
to follow the process of what happens with the truck after it leaves the building. Furthermore,
we believe that the future with AI will allow for better forecasting, optimization of transport
solutions, identification of disruptions in data analysis instead of manually. It will be a large
difference, might not today, but rather in the future.
5. Discussion
This chapter will discuss the applicability of the various industry 4.0 technologies on Volvo
AB’s supply chain.
Volvo AB integrates machine learning into its supply chain to detect patterns within extensive
datasets and predict future trends. In terms of risk management, Volvo AB employs the AI
solution Prewave to scan newspapers for indications of potential disruptions or risks.
Operating as a global entity, Volvo AB adheres to stringent policies and governance on
information sharing, influencing the quality and accuracy of data fed into AI algorithms
(Barghava et al., 2022). The system has been supplied with data concerning supplier names
and addresses, which could suffice for scanning newspapers and media for indications of
potential disruptions. However, enhancing the input of qualitative data, such as historical
records of successful on-time deliveries, quality concerns, and supplier responsiveness to
sudden changes or requirements, could significantly improve the AI’s ability to identify
disruptions or risks. This augmentation would enable more accurate indications and enhance
proactive risk management, due to the quality and accuracy of AI predictions being
dependent upon the input data quality (Barghava et al., 2022). AI holds the potential to
enhance Volvo AB’s Enterprise Risk Management (ERM) system employed to identify,
mitigate and report potential risks. ERM systems can be integrated with AI, presenting a
promising solution to enhance incident prediction and risk assessment automation. It enables
efficient analysis of vast datasets, providing insights that contribute to a more comprehensive
understanding of risks within the organization and offers the potential to identify
dependencies and correlations between risks (Sharma, 2019). If integrated with Volvo AB’s
ERM system, it allows for not only identification and categorization of potential disruptions
and risks, but identification of the dependencies and correlations between them.
Moreover, Volvo AB’s extensive network of suppliers introduces complexities in
collaboration and data visibility, compounded by stakeholders’ security concerns.
Consequently, larger enterprises with limited data resources tend to encounter fewer issues, as
AI models benefit from industry scale, whereas smaller businesses face challenges and
53
implications due to data limitations (Barghava et al., 2022). Volvo AB’s strict governance on
information sharing and stakeholders’ concern for safeguarding against cyberthreats pose
challenges and complexities in fully leveraging the potential of Prewave. It could be in Volvo
AB’s interest to establish communication among the stakeholders, to reduce uncertainties that
can arise due to sharing barriers (Afrifah et al., 2022).
Additionally, Cloud Computing holds promise for Volvo AB in enhancing resilience by
swiftly identifying and mitigating disruptions, thereby assisting downstream suppliers in
preparing contingency measures (Chen & Chang, 2021). Integration of cloud computing
enhances structural flexibility, responsiveness and ultimately fosters competitive advantages
(Gammelgaard & Nowkicka, 2023). Given Volvo AB’s extensive supplier base, effective risk
identification and mitigation can be considered crucial for minimizing disruptions throughout
the supply chain. However, the successful adoption of cloud computing requires stakeholders
to understand its architecture, characteristics, and interrelationships to establish robust
protection mechanisms against disruptions and risks (Herrera & Janczewski, 2015).
Therefore, it is essential that all stakeholders within Volvo AB’s supply chain possess
awareness and knowledge of cloud computing for its successful utilization. This will become
resource-intensive and complex due to the extensive supply network.
Cloud computing entails a shared pool of IT resources, enabling cost reduction, scalability,
and rapid development, thereby enhancing agility, flexibility, and efficiency (Giannakis et al.,
2019). Despite its immense advantages, widespread adoption within Volvo AB’s supply chain
necessitates trust among stakeholders, as it increases vulnerability to cybersecurity threats
(Chen & Chang, 2021). Disclosing information within the supply chain poses risks including
potential exposure of competitive advantage or vulnerabilities, association with unfavorable
supplier practices or suspicion from consumers about hidden information (Sodhi et al., 2019).
Such leakage of data and information from Volvo AB could cause scandals, jeopardizing their
position in the market and customer loyalty. Convincing all stakeholders to embrace this
technology and willingly share resources and information poses a complex challenge,
especially considering the multitude of suppliers involved. Smaller enterprises may find
accessing a shared IT pool more profitable due to resource constraints, while Volvo AB, with
its ample resources, may opt to develop in-house unique solutions and market them to
customers for competitive advantage and profitability.
Volvo AB has strategically integrated Internet of Things technology within its trucks,
launching smarter connectivity and predictive maintenance capabilities (Personal
Communication, 2024). An implementation of IoT is aimed at enhancing efficiency, reducing
costs, ensuring product quality and fostering sustainability within a company’s operations
(Sallam et al., 2023). IoT's broader impact on the supply chain is profound, providing
real-time temporal and spatial insights into product flows, which aids in the swift detection
and resolution of inconsistencies (Ben-Daya et al., 2022). This adoption across Volvo AB’s
supply chain could enhance real-time monitoring of product flows, enabling proactive
interventions in the face of disruptions or risks. Such proactive measures not only mitigate
potential costs and risks but also expedite recovery processes, ensuring efficiency and
54
resilience. Furthermore, IoT’s analytical ability extends to historical and real-time trends
analysis, improving demand predictions, refining accuracy and enhancing forecasting
capabilities (Sallam et al., 2023). By harnessing these capabilities, Volvo AB can enhance its
proactive stance, leveraging IoT to analyze historical data on key product flows and detect
patterns indicative of potential disruptions.
However, it is imperative to acknowledge the complexities and difficulties surrounding IoT,
particularly concerning data security and interoperability issues (Sallam et al., 2023).
Operating as a global entity, Volvo AB places the utmost importance on data security,
necessitating stringent measures to safeguard sensitive information. Moreover, within supply
chains, legacy systems often prevail, posing compatibility challenges for IoT integration
(Sallam et al., 2023). Given Volvo AB’s extensive supply chain with numerous tiers, ensuring
supplier compliance and system integration becomes resource-intensive, complex and costly.
Addressing the absence of global standards further compounds integration difficulties,
leading to expensive and labor-intensive processes (Sallam et al., 2023). To mitigate these
challenges, Volvo AB must invest resources in expertise, foster communication and broker
agreements with suppliers to establish standards and regulations. This will foster
standardization within their supply chain, which enhances data protection among
stakeholders.
In addition to the IoT, Volvo AB could explore the adoption of Artificial Intelligence of
Things (AIoT) and Explainable Artificial Intelligence (XAI). AIoT promises smarter business
decisions, improved operational efficiency and the elimination of irrelevant data, culminating
in more accurate decision-making (Aliahmadi et al., 2022). By integrating AIoT into its
supply chain, Volvo AB stands to streamline processes, reduce costs, and enhance
decision-making efficacy through automated data analysis and real-time assessment. Despite
its potential, AIoT presents challenges akin to IoT, necessitating supplier collaboration and
difficulties with computational complexities across diverse resources (Zhang & Tao, 2021).
Given its global footprint, Volvo AB recognizes the significance of informed
decision-making, underlining the potential value of XAI in deepening the understanding of
AI-driven outcomes (Nimmy et al., 2022). XAI provides enhanced explainability,
transparency and expedited adoption, fostering trust and understanding of AI within the
organization (Kangra & Singh, 2022). This transparency aids in overcoming adoption
hurdles, such as ineffective change management and vague understanding of Industry 4.0
technologies (Cannas et al., 2023), thereby facilitating smoother integration and optimization
operations for Volvo AB. Accurate communication among stakeholders emerges as pivotal
for optimizing operations, enhancing performance, agility and adaptability within the supply
chain (Afrifah et al., 2022).
Artificial Neural Networks are the most adopted AI technology within supply chain
management, due to its ability in discerning, understanding and predicting patterns within
datasets beyond human capability. This enables Volvo AB to focus their proactive efforts on
the variables with the greatest influence, thereby mitigating disruptions and reducing
potential impacts more efficiently, rather than risking diversion of attention to less pertinent
55
factors. ANN excels both in standalone applications and when combined with other AI
technologies, yielding improved accuracy and performance (Toorajipour et al., 2021). This
adaptability renders ANNs particularly pertinent to Volvo AB, especially in conjunction with
Prewave. A case study conducted by Oliveira et al. (2013) presents the efficacy of ANNs in
forecasting stock prices within the finance market, outperforming conventional methods by
discerning behavioral trends with accuracy. Such predictive capabilities hold promise for
Volvo AB, enabling proactive measures to address potential price fluctuations by selecting
alternative suppliers.
Moreover, ANNs provides a powerful tool for analyzing supplier performance metrics,
including quality, delivery times and pricing (Soori et al., 2023). This could provide Volvo
AB with a comprehensive framework to assess supplier reliability and efficacy. Given
Volvo’s extensive network of 1500 suppliers, diligent monitoring and evaluation are
imperative to uphold a seamless product flow, a task where ANNs excel by flagging
deviations and altering stakeholders to secure timely deliveries (Soori et al., 2023). Notably,
another case study revealed that ANN compared to traditional strategies, provides more
accurate forecasts (Seyedan & Mafakheri, 2020), further underlining the merit of its
integration within Volvo AB.
Volvo AB needs to carefully weigh the implications of implementing ANNs technology.
ANN entails managing diverse data issues, such as imbalanced data, high dimensionality, and
limited data for learning (Prieto et al., 2016). This highlights the interplay with other
technologies, including data sharing and accessibility of information. Optimal functionality
demands substantial data volumes, which can pose a resource-intensive endeavor (Richey et
al., 2023). It is worth noting that ANN excels when tackling problems with simple input and
output data, as larger datasets may introduce complexities (Praveen et al., 2019). Hence,
Volvo AB must exercise caution to avoid misidentifying problems suitable for ANN
solutions, as this could significantly impact implementation success (Shirvastav, 2022).
Agent-based systems have found widespread application across various domains within
supply chain management, including supply chain planning, system simulation, and analysis
of complex behaviors (Toorajipour et al., 2020). At Volvo AB, an implementation of
Multi-Agent Systems (MAS) holds particular relevance. MAS comprises multiple interacting
agents, which may be homogenous or heterogenous in nature (Kaur & Kaur, 2022). These
agents represent various actors within the supply chain, such as suppliers, customers, and
manufacturers. The system dynamically adapts their behavior to anticipate, respond to,
enhance flexibility, and recover from both anticipated and unforeseen disruptions when
simulating diverse scenarios (Kassa et al., 2023). Currently, Volvo AB employs an Enterprise
Risk Management (ERM) system, which serves to report, review, mitigate and identify risks
(Personal Communication, 2024). However, an alternative solution proposed by Sohbrabi et
al. (2018) introduces an AI-driven scenario planning approach integrated within an ERM
framework. This solution, termed Scenario Planning Adviser (SPA), leverages AI techniques
to generate potential scenarios by analyzing raw data from sources including news and social
media posts. These scenarios not only depict the current situation but also forecast potential
56
future effects of observed phenomena or identified key risk drivers. By integrating MAS as
an AI scenario planning tool, Volvo AB can gain insights into stakeholder behavior, thereby
enhancing their ERM capabilities in risk aggregation and facilitating the formulation of
proactive action plans. This approach enables the organization to anticipate and respond
effectively to emerging risks, thereby enhancing its resilience in an increasingly dynamic
business environment.
Information sharing and data privacy pose significant concerns for the supply chain and its
stakeholders, often constraining and limiting the adoption of processes and technologies.
Blockchain technology is promoted for its potential to enhance trust among actors involved in
information and data exchange, as it enables systems to communicate directly with each other
(Longo et al., 2019). In the realm of risk management, blockchain plays a crucial role in
managing and mitigating risks by preventing security breaches and enhancing connectivity
(Min, 2019). Furthermore, the implementation of blockchain holds promise for enhancing
operational efficiency within the supply chain, leveraging its features of transparency,
traceability and decentralization (Lai et al., 2021). Consequently, integrating blockchain into
Volvo AB’s supply chain could enhance overall performance and improve visibility,
resilience and transparency.
However, the adoption of blockchain necessitates agreement among stakeholders and
adequate resources for installation. Presently, due to existing contractual requirements and
agreements, Volvo AB lacks the requisite conditions and possibilities for such
implementation. Instead, the focus is on evaluating enhanced visibility and resilience through
the development of Prewave utilization (Personal Communication, 2024). In scenarios where
contractual barriers are absent, implementing blockchain could improve resilience within
Volvo AB’s supply chain by fostering improved collaboration among stakeholders, thereby
enhancing transparency and trust (Lohmer et al., 2020). Moreover, this technology has the
potential to mitigate risks associated with supply and demand fluctuations, behavioral
uncertainties and vulnerabilities in information security, including fraud (Alkhudary et al.,
2022). This becomes particularly relevant for a global entity like Volvo AB, where data
breaches could have profound organizational consequences.
6. Analysis
This chapter will analyse the AI technologies in Volvo AB’s Supply Chain and the different
application areas, that will be analysed both in terms of challenges and risks regarding the
implementation but also future AI possibilities for Volvo AB.
6.1. AI technologies in Volvo AB’s Supply Chain
Volvo AB operates within a complex global supply chain environment where effective risk
management and technological innovation are paramount. The integration of various
technologies such as ANN, Cloud Computing, IoT, and Blockchain offer promising solutions
to enhance resilience, efficiency, and transparency throughout the supply chain (Toorajipour
57
et al., 2021; Chen & Chang, 2021; Sallam et al., 2023; Alkhudary et al., 2022). ANN emerge
as the most suitable AI technology for Volvo AB’s supply chain management. ANNs excel in
discerning, understanding and predicting patterns within datasets, enabling proactive risk and
disruption mitigation (Toorajipour et al., 2021). Moreover, ANNs provide a robust framework
for analyzing supplier performance metrics, facilitating informed decision-making and
ensuring product flow efficiency (Soori et al., 2023). However, the implementation of ANNs
requires careful consideration of data quality, resource availability and potential complexities
associated with large datasets (Richey et al., 2023; Praveen et al., 2019). To overcome these
challenges, Volvo AB should prioritize change management, emphasizing effective
organizational practices and collaboration among stakeholders.
In addition to ANNs, other technologies such as Cloud Computing, IoT and Blockchain offer
valuable opportunities to enhance resilience and transparency within Volvo AB’s supply
chain. Cloud Computing can enhance structural flexibility and responsiveness (Gammelgaard
& Nowkicka, 2023), while IoT enables real-time monitoring of product flows and predictive
maintenance capabilities (Sallam et al., 2023). Blockchain technology holds promise for
improving transparency and trust among supply chain stakeholders (Longo et al., 2019),
although its implementation requires overcoming contractual barriers and resource limitations
for Volvo (Personal Communication, 2024). Furthermore, Agent-based systems could
potentially be a profitable decision to implement for Volvo AB due to its capabilities of
enhancing ERM (Sohrabi et al., 2018). Implementing ANN instead of an Agent-based
system could be advantageous for Volvo AB due to several reasons. ANNs excels at pattern
recognition and prediction, making them valuable for identifying trends and forecasting
outcomes in the supply chain (Toorajipour et al, 2021). ANN further offers tools for
analyzing supplier performance metrics, aiding in supplier evaluation and selection (Soori et
al., 2023). It outperforms traditional methods in terms of accuracy and performance (Seyedan
& Mafakheri, 2020; Oliveira et al., 2013), providing Volvo AB with better-informed
decision-making and risk management strategies. Additionally, the broad applicability of
ANNs across industries suggest seamless integration into Volvo AB’s operations, providing
resource efficiency and scalability. Overall, Volvo AB must strategically evaluate these
technologies, considering their respective advantages, implications and alignment with
organizational goals and values.
6.2. AI in different application areas for Volvo AB’s supply chain
Within supply chains, particularly in the field of forecasting, incorporating AI tools is crucial
for organizations when aiming to bolster decision-making capabilities and mitigate risks
(Seyeden & Mafakheri, 2020). As highlighted, Volvo AB has implemented a centralized
Enterprise Risk Management system that emphasizes their dedication to systematic risk
assessment and mitigation strategies across various domains (Volvo AB, 2023). This
approach emphasizes proactive risk identification and mitigation within the supply chain
operations, and aligns with contemporary literature. To enhance supply chain resilience and
enhance decision-making that are driven by big data analytics, the adoption of intelligent
forecasting techniques is paramount. According to Seyeden and Mafakheri (2020) benefits of
58
data-driven forecasting methods are enhanced operational efficiency due to the more accurate
estimations provided. Additionally, the significance of advanced analytics in forecasting for
mitigating supply chain risks and optimizing performance has been highlighted (Pournader et
al., 2021). By leveraging AI tools for forecasting models, organizations can anticipate
demand fluctuations, and by that optimize inventory management practices and control, and
proactively address potential risks connected to the supply chain (Riahi et al., 2021).
Essentially, the integration of intelligent forecasting and robust risk management practices
underscores the importance of leveraging data-driven insights in order to navigate risks
(Sharma et al., 2022). This strategic alignment between advanced forecasting techniques and
structured risk management frameworks exemplifies Volvo AB's proactive approach to
driving operational efficiency and resilience within their supply chain network.
Another proactive approach that Volvo AB demonstrates to mitigate risks and enhance
operational resilience is by the integration of an external AI tool that scans global media
sources for potential crises and disruptions related to their network of suppliers. This
addresses and identifies vulnerabilities within their supply chain of network (Personal
Communication, January 2024). Advantages of using AI in procurement as highlighted by
Meyer & Henke (2023) aligns well with Volvo AB’s proactive approach to manage risks. The
integration of AI in procurement automates operational tasks, supports employees and
provides data-driven outcomes of information which enhance the performance of Volvo AB’s
procurement department. Therefore, leading to competitive advantage by its effective way of
handling large amount of data shared between the suppliers and the buyers (Meyer & Henke,
2023). By applying the AI-driven external tool, a risk management culture among all actors is
essential and a way to enhance resilience in the supply chain. This can be accomplished by
comprehensive understanding of the network's structure, in which Volvo AB is pursuing this
approach by close communication with suppliers and set criteria on their suppliers (Personal
Communication, January 2024). This proactive commitment aims to mitigate risks and
disruptions within Volvo Ab’s supply chain.
6.3. AI implementation Risks and Strategy for Volvo AB
Challenges and barriers in AI implementation encompass concerns such as data privacy,
ethics, and transparency (Richey et al., 2023). AI heavily relies on computer software
infrastructure (Min, 2010), and its successful integration hinges on accessing vast amounts of
suitable data alongside the right expertise and competencies (Lynn et al., 2019). However,
this reliance on extensive data presents risks concerning data privacy, as AI necessitates
significant data volumes for optimal functionality. Any inaccuracies or gaps in data, as well
as potential corruption, can render AI models ineffective, impeding the realization of their
full potential (Richey et al., 2023). Currently, Volvo AB is feeding its AI solution, Prewave,
with supplier data, underscoring the need to manage data security and privacy regulations
effectively to expand its data integration responsibly. Furthermore, operating in a legal gray
area, AI poses challenges regarding regulatory compliance, accountability, and trust (Dogru
& Keskin, 2020). Volvo AB must navigate this dynamic landscape collaboratively with its
suppliers to establish a clear regulatory framework for AI utilization.
59
As a global enterprise, Volvo AB encounters organizational and operational barriers in AI
implementation. Organizational barriers include a lack of AI understanding, unclear project
goals, and misaligned strategies among supply chain stakeholders (Shirvastav, 2022). Volvo
AB should acknowledge the need for effective change management, prioritizing
organizational and developmental practices to adapt AI to its supply chain effectively (Hangl
et al., 2022; Personal Communication, 2024). Success hinges on top management support,
which relies on fostering trust in AI, accurately identifying problems suitable for AI
solutions, and committing to AI initiatives (Shirvastav, 2022). Resistance to new technology
is common within organizations, necessitating transparent and inclusive management to
ensure successful AI integration (Hangl et al., 2022). Volvo AB must cultivate transparency,
commitment and employee engagement to foster acceptance and mitigate internal resistance
effectively.
In leveraging AI, organizations should focus on optimizing existing processes and aligning
AI priorities with business objectives (Kruhse-Lehtonen & Hofmann, 2020). Volvo AB must
identify AI-compatible problems aligned with its core values and processes to maximize
advantages while considering resource demands, capital requirements and inherent risks
(Richey et al., 2023). Infrastructure refinement and collaborative partnerships among supply
chain actors are vital for fostering AI implementation (Cannas et al., 2023). Due to Volvo
AB’s extensive supply chain network, it is perceived too complex to alter the supply chain for
AI implementation rather than adjusting the AI to be suitable for the supply chain (Personal
Communication 2024). Volvo AB should therefore prioritize a committed leadership and
alignment with business priorities when integrating AI.
Helo & Hao (2022) propose an AI implementation approach focusing on data assessment and
preparation, involving data collection, organization and structured labeling to enhance AI
learning. This approach involves two primary steps: firstly gathering data from industrial
sensors and IoT systems and then organizing and structuring this data. This process will
effectively eliminate duplicates, manage missing data attributes and appropriately label data
essential for AI’s learning process. The second step emphasizes the learning and training
aspect of AI through iterative learning cycles, learning to for instance enhanced forecasting
accuracy, regardless of the dataset's size. In the context of Volvo AB, this approach could be
applied by evaluating data from their IoT systems integrated into their trucks, alongside data
from other sensors and their Prewave AI solution. However, leveraging the learning aspect of
AI necessitates substantial resources and expertise for training, which may pose challenges in
terms of labor and time intensiveness. Therefore, while promising, the implementation of this
approach would require careful consideration of resource allocation and skill development
within the organization.
To effectively mitigate the risks associated with implementing AI, Volvo AB should embrace
a comprehensive strategy. This strategy must begin by establishing a clear regulatory
framework in collaboration with both suppliers and regulatory authorities. By doing so, Volvo
AB can ensure compliance with data privacy regulations and proactively address any legal
60
uncertainties that may arise. Additionally, Volvo AB should prioritize investment in change
management initiatives to overcome organizational barriers. This includes educating
employees about AI, establishing transparent project goals, and aligning strategies across all
supply chain stakeholders. Committing to resource allocation, both in terms of labor and
infrastructure, is essential to support AI implementation effectively. This may necessitate
investment in training programs to develop the necessary expertise and optimizing existing
infrastructure to seamlessly integrate AI.
Furthermore, fostering AI acceptance within the organization requires cultivating
transparency, commitment, and employee engagement. Establishing open communication
channels to address concerns and effectively mitigate internal resistance will facilitate a
smooth implementation of AI initiatives within Volvo AB. Finally, they could benefit from
embracing AI implementation approaches that prioritize data assessment, preparation, and
iterative learning cycles, as suggested by Helo & Hao (2022). This approach enables Volvo
AB to enhance AI capabilities while efficiently addressing resource constraints through
continuous learning processes. By adopting these strategies, Volvo AB can effectively
navigate the inherent risks associated with AI implementation and maximize the advantages
of AI integration within their operations.
6.4. AI possibilities for Volvo AB in the future
Volvo AB’s future with AI holds multifaceted opportunities. One of the opportunities lies in
leveraging AI-driven forecasting tools within their supply chain operations, where Riahi et al.
(2021) highlighted the potential of AI usage in demand forecasting, an area for Volvo AB’s
market responsiveness and customer satisfaction. Furthermore, utilizing Big data analytics
gives Volvo AB the opportunity to explore supervised and unsupervised learning approaches
for demand forecasting, where Volvo AB can analyze both historical data and by that
generate more precise demand estimations and identification of nuanced patterns within
datasets (Seydan & Mafakheri, 2020). Second opportunity for Volvo AB is to adapt
machine-learning-driven inventory management which enables the organization to swiftly
adjust to fluctuating market demands and supply chain disruption, ensuring robustness in
inventory planning. This enhances not only the operational flexibility but also reduces related
costs (Ahmadi et al., 2022).
Third possibility for Volvo AB is to optimize resource allocation across the network and
enhance resilience by an examination of disruption effects on individual suppliers in the
supply chain, where it according to (Li et al. (2021) enables the identification of areas of
vulnerability and relocates resources to the needed area in case of disruption. Leading to the
outcome of where Volvo AB possibly can increase its competitiveness and ensure its ability
to grow in such a complex supply chain environment. The fourth possibility for Volvo AB
lies in the integration of AI in procurement and supply chain management, a transformation
of shifting the focus from operational to strategic aspects with improvements in deeper
understanding in various factors. By the usage of AI, Volvo AB can analyze vast amounts of
data to uncover insights that lead to enhanced decision-making and strategic planning which
61
leads to minimizing risks (Allal-Chérif et al., 2021). The importance lies in providing AI with
all the data needed, including the supplier's data that can be shared across other tiers in the
supply chain network (Chopra, 2019). Volvo AB acknowledges the challenge and the
importance of the quality of data, its availability and relevance (Personal Communication
2024), and by doing so Volvo can ensure to fully realize the capabilities of AI in
procurement.
Volvo AB holds a strong interest in the enhancement of forecasting as a future goal, while
underscoring the supply chain visibility as a pivotal and interesting area. Additionally, they
are convinced that employing data analysis for identifying disruptions instead of through
manual methods will lead to significant future improvements (Personal Communication
2024). An another potential possibility that Volvo AB have with AI in the future are within
supply chain regulatory and compliance risks in which according to Sodhi et al. (2019) it
improves the visibility and transparency in the supply chain by emphasizing the importance
of the advancing of environmental and social sustainability initiatives without disclosing
sensitive and competitive information, which can be challenging. However, by doing so,
Volvo AB strikes a balance between transparency and risk mitigation related to the reputation
of the organization. Yet, according to Sabia (2019) applying AI within corporate compliance
through big data analysis and real-time risk alerts, emerges as an opportunity for Volvo AB.
Furthermore, it has potential in exploring the possibilities of AI in the future regarding
resilience planning and strategy. Specifically by adopting proactive approaches supported by
AI-driven tools to explore various resilience strategies to understand the scenario and to
identify the most effective actions needed. Meaning that, it allows Volvo AB to enhance
decision-making on how to prepare and respond to supply chain disruptions (Baryannis et al.,
2019). Lastly, there are potential possibilities for Volvo AB to integrate AI-driven tools in
supplier risk management where they can optimize its supplier selection processes by setting
needed criteria (Sahoo et al., 2024). Integrating AI into Volvo AB’s supplier selection and
risk assessment processes would not only enhance efficiency but also provide an opportunity
to uphold their specific and certain requirements that their suppliers need to fulfill (Personal
Communication 2024).
7. Conclusions
This chapter will conclude the research findings, acknowledge limitations and suggest
directions for future research.
In conclusion, as Volvo AB delves into the realm of AI implementation within their complex
global supply chain, they must read the future carefully, recognizing both the opportunities
and challenges that lie ahead. The integration of advanced technologies such as ANN, Cloud
Computing, IoT and Blockchain holds great promise in enhancing resilience, efficiency and
transparency throughout Volvo AB’s supply chain. While AI, particularly ANN, emerges as a
powerful tool for discerning patterns, predicting outcomes, and optimizing processes, its
successful implementation requires a comprehensive strategy. This strategy must encompass
62
establishing clear regulatory frameworks to ensure compliance with data privacy regulations,
prioritizing change management initiatives to overcome organizational barriers, and
committing to resource allocation for effective AI integration.
Volvo AB stands at the forefront of innovation and adaptation in the industry with diverse
opportunities and possibilities to leverage Artificial Intelligence to advance across various
domains, possibilities such as enhanced efficiency, mitigation of risks and enhanced
competitive advantages. The integration of AI enables Volvo AB to make data-driven
decisions, anticipate risks and adapt proactively to fluctuations. Moreover, fostering a culture
of transparency, commitment and employee engagement will be pivotal in driving acceptance
and mitigating internal resistance to AI initiatives within Volvo AB. Embracing AI
implementation approaches that emphasize data assessment, preparation and iterative
learning cycles can further enhance Volvo AB’s capabilities while addressing resource
constraints. Ultimately, by adopting a strategic and holistic approach to AI implementation,
Volvo AB can navigate the inherent risks and maximize the advantages of AI integration
within their operations, thereby solidifying their position as a leader in the ever-evolving
landscape of supply chain management.
7.1. Limitations & Future Research
This research has various limitations identified as confidentiality, generalizability, scope and
bias. Volvo AB may consider some data and information too sensitive and confidential to
disclose publicly, affecting the depth and insights of the study. Findings of the study may not
be considered generalizable to other companies or industries due to the specific context of
Volvo’s operations and the unique tools, strategies and challenges associated with them.
Limitations related to the chosen methodology of the interviews and the restricted scope of
only two interviewees with similar roles, may lead to overlooking important insights and
might not capture a wide diversity of opinions and experience within the organization. The
interviews provide valuable information, but the necessity remains to cautiously draw
conclusions in the report solely based on this limited data. Lastly, potential bias could be
present in the data and selection of sources, which may impact the reliability and validity of
the research findings.
There are several directions that future research could explore. It could apply a broader
industry perspective and conduct research that explores the application of AI in risk
management beyond just internal operations. This could involve studying risk factors specific
to different industries and revealing new insights into effective risk management strategies.
Another suggested direction could be to investigate the classification of supply chain risks to
improve the understanding of their nature and implications. Such research could aid in
developing targeted risk mitigation strategies tailored to different risk categories.
Furthermore, future research could explore how various AI technologies can facilitate
dynamic resource reorganization to address supply chain risks effectively. This could involve
studying the nature of data required for dynamic decision making and the AI algorithms that
support real-time resource allocation.
63
Analyzing potential barriers to AI adoption in procurement departments, such as competency
gaps, cultural resistance and organizational maturity, could be a suggestion for future
research. Such study could provide insights into overcoming these obstacles and accelerating
AI adoption within procurement functions. Lastly, future research can focus on exploring the
robustness and interpretability of AI algorithms in supply chain management. The focus of
the study could be on developing transparent and reliable AI models that can easily be
interpreted and trusted by stakeholders within the supply chain.
The significance of this study could be improved through future research. A suggestion is to
involve key stakeholders from Volvo AB and other relevant organizations in the research
process to ensure their perspectives and insights are incorporated into the study. Engaging
with industry experts, practitioners and policymakers would validate the research findings
and enhance its relevance and applicability.
64
References
Aboutorab, H., Hussain, O.K., Saberi, M., Khadeer, K.H., & Prior, D. (2023). Adaptive
Identification of Supply Chain Disruptions through Reinforcement Learning. Expert Systems
with applications. https://doi.org/10.1016/j.eswa.2024.123477
Adeoye-Olatunde, O.A., & Olenik, N.L. (2021). Research and scholarly methods:
Semi-structured interviews. Journal of the American college of clinical pharmacy, 4(10), p.
1358-1367. https://doi.org/10.1002/jac5.1441
Adeyemi, O.A., Ayoade, I.A., Adeyemi, O.E., Olanrele, O.O., Eweoya, I.O., & Matthew, J.
(2023). Adoption of Artificial Intelligence in Operations Management: A design-oriented
Review. Advances in Manufacturing Technology. DOI: 10.3233/ATDE230892
Afrifah, Epiphaniou, G., Ersotelos, N., & Maple, C. (2022). Barriers and Opportunities In
Cyber Risk And Compliance Management For Data-Driven Supply Chains. Annual Hawaii
International Conference on System Sciences, p. 1812–1821.
https://doi.org/10.24251/hicss.2022.225
Ahmadi, Mosadegh, H., Maihami, R., Ghalehkhondabi, I., Sun, M., & Süer, G. A. (2022).
Intelligent inventory management approaches for perishable pharmaceutical products in a
healthcare supply chain. Computers & Operations Research, 147, 105968.
https://doi.org/10.1016/j.cor.2022.105968
Ahmarofi, A.A., Ramli, R., & Abidin, N.Z. (2017). Predicting Completion Time for
Production Line in a Supply Chain System through Artificial Neural Networks. Journal of
Supply Chain Management, 6(3).
Aliahmadi, A., Nozari, H., & Ghahremani-Nahr, J. (2022). AIoT-based Sustainable Smart
Supply Chain Framework. International Journal of Innovation in Management Economics
and Social Sciences, 2(2), p. 28-32. https://doi.org/10.52547/ijimes.2.2.28
Alkhudary, R., Queiroz, M.M., & Fenies, P. (2022). Mitigating the risk of specific supply
chain disruptions through blockchain technology. Supply Chain Forum: An International
Journal, 25(1), pp. 1-11. https://doi.org/10.1080/16258312.2022.2090273
Allal-Chérif, O., Simon-Moya, V., Cuenca Ballester, A.C. (2021). Intelligent purchasing:
How artificial intelligence can redefine the purchasing function. Journal of Business
Research, 124, p. 69-76. https://doi.org/10.1016/j.jbusres.2020.11.050
65
Angkiriwang, Pujawan, I. N., & Santosa, B. (2014). Managing uncertainty through supply
chain flexibility: reactive vs. proactive approaches. Production & Manufacturing Research,
2(1), 50–70. https://doi.org/10.1080/21693277.2014.882804
Atwani, M., Laaziz, E.H., Hlyal, M., & Alami, J.E. (2020). AI in Supply chain performance
and resilience: a literature review. International Conference on Modelling, Optimization and
SIMulation, 13.
Azzi, R., Chamoun, R.K., & Sokhn, M. (2019). The power of a blockchain-based supply
chain. Computers & Industrial Engineering, 135, pp. 582-592.
https://doi.org/10.1016/j.cie.2019.06.042
Barrad, S. (2020). Impact of Artificial Intelligence, Analytics, And Procurement Strategy on
Cost Reduction (Quebec Studies in Information Science and Technology). [Doctoral thesis].
University of Quebec. Assessed 04-11 from Sherif_Barrad_2020_these.pdf (uqo.ca)
Barroso, A.P, Machado, V.H., & Machado, V.C. (2008). A supply chain disturbances
classification. IEEE International Conference on Industrial Engineering and Engineering
Management, pp. 1870-1874. doi: 10.1109/IEEM.2008.4738196.
Baryannis, G., Dani, S., & Antoniou, G. (2019). Predicting supply chain risks using machine
learning;: The trade-off between performance and interpretability. Future Generation
Computer Systems, 101, p. 993-1004. . https://doi.org/10.1016/j.future.2019.07.059
Baryannis, Validi, S., Dani, S., & Antoniou, G. (2019). Supply chain risk management and
artificial intelligence: state of the art and future research directions. International Journal of
Production Research, 57(7), 2179–2202. https://doi.org/10.1080/00207543.2018.1530476
Belhadi, A., Mani, V., & Kamble, S.S., & Rehman Khan, S.A. (2021). Artificial
intelligence-driven innovation for enhancing supply chain resilience and performance under
the effect of supply chain dynamism: an empirical investigation. Annals of Operation
Research. https://doi.org/10.1007/s10479-021-03956-x
Belhadi, A., Kamble, S., Wamba, S.F., Queiros, M.M. (2021). Building supply-chain
resilience: an artificial intelligence-based technique and decision-making framework.
International Journal of Production Research, 60(22), p. 4487-4507.
https://doi.org/10.1080/00207543.2021.1950935
Ben-Daya, M., Hassini, E., & Bahroun, Z. (2022). A Conceptual Framework for
Understanding the Impact of the Internet of Things on Supply Chain Management.
Operations and Supply Chain Management, 15(2), p. 251-268.
http://doi.org/10.31387/oscm0490345
66
Bengtsson, M. (2016). How to plan and perform a qualitative study using content analysis.
NursingPlus Open, 2, p. 8-14. https://doi.org/10.1016/j.npls.2016.01.001
Bhargava, A., Bhargava, D., Kumar, P. N., Sajja, G. S., & Ray, S. (2022). Industrial IoT and
AI implementation in vehicular logistics and supply chain management for vehicle mediated
transportation systems. International Journal of System Assurance Engineering and
Management, 13(1), p. 673-680.
Bhatti, & Bauirzhanovna, B. A. (2023). IMPACT OF INTELLIGENT INVENTORY
SYSTEM ON IMPROVEMENT OF REVERSE LOGISTICS: A CASE OF SAUDI
MANUFACTURING INDUSTRY. Operational Research in Engineering Sciences: Theory
and Applications, 6(1), 1–19. https://doi.org/10.31181/oresta/060101
Blos, M.F., Da Silva, R.M., & Miyagi, P.E. (2015). Application of an Agent-based Supply
Chain to Mitigate Supply Chain Disruptions. IFAC-PapersOnline, pp. 640-645.
Bode, C., & Wagner, S.M. (2014). Structural drivers of upstream supply chain complexity
and the frequency of supply chain disruptions. Journal of Operations Management, 36, pp.
215-228. https://doi.org/10.1016/j.jom.2014.12.004
Bodendorf, F., Sauter, M., & Franke, J. (2023). A mixed methods approach to analyze and
predict supply disruptions by combining causal inference and deep learning. International
Journal of Production Economics, 256. https://doi.org/10.1016/j.ijpe.2022.108708
Brintrup, A., Kosasih, E., Schaeffer, P., Zheng, G., Demirel, G., & MacCarthy, B.L. (2023).
Digital supply chain surveillance using artificial intelligence: definitions, opportunities and
risks. International Journal of Production Research.
https://doi.org/10.1080/00207543.2023.2270719
Brintrup, A., Pak, J., Ratiney, D., Pearce, T., Wichmann, P., Woodall, O., & McFarlane, D.
(2019). Supply chain data analytics for predicting supplier disruptions: a case study in
complex asset manufacturing. International Journal of Production Research, 58(11).
https://doi.org/10.1080/00207543.2019.1685705
Brockmann, N., Kosasih, E.E., & Brintruo, A. (2022). Supply Chain Link Prediction on
Uncertain Knowledge Graph. ACM SIGKDD Explorations Newsletter, 24(2), p. 124-130.
https://doi.org/10.1145/3575637.3575655
Bromiley, P., McShane, M., Nair, A., & Rumstambekov, E. (2015). Enterprise Risk
Management: Review, Critique, and Research Directions. Long Range Planning, 48(4), p.
265-276. https://doi.org/10.1016/j.lrp.2014.07.005
Browning, T., Kumar, M., Sanders, N., ManMohan, S.S., Thurer, M., & Tortorella, G.L.
(2023). From supply chain risk to system-wide disruptions; research opportunities in
67
forecasting, risk management and product design. International Journal of Operations &
Production Management, 43(12).
Bryman, A., & Bell, E. (2011). Företagsekonomiska forskningsmetoder. Liber AB.
Cannas. V.G., Ciano, M.P., Saltalamacchia, M., & Secchi, R. (2023). Artificial Intelligence in
supply chain and operations management: a multiple case study research. International
Journal of Production Research. https://doi.org/10.1080/00207543.2023.2232050
Chang, S.E., & Chen, Y. (2020). When Blockchain Meets Supply Chain: A Systematic
Literature Review on Current Development and Potential Applications. IEEE Access, 8, pp.
62478-62494. doi: 10.1109/ACCESS.2020.2983601
Chen, M.L, & Chang, W.L. (2021). Supply - and cyber - related disruption in cloud supply
chain firms: Determining the best recovery speeds. Transportation and Research Part E:
Logistics and Transportation Review, 151. https://doi.org/10.1016/j.tre.2021.102347
Chopra, A. (2019). AI in Supply & Procurement. Amity International Conference on
Artificial Intelligence (AICAI), p. 308-316. DOI: 10.1109/AICAI.2019.8701357.
Chopra, S., & Sodhi, M. (2014). Reducing the Risk of Supply Chain Disruptions. MIT Sloan
Management Review Magazine Spring.Reducing the Risk of Supply Chain Disruptions
(mit.edu)
Collis, J., & Hussey, R. (2021). Business research: A practical guide for students. RED
GLOBE PRESS.
Darnall, N., Welch, E.W., & Cho, S.K. (2019). Chapter 30: Sustainable supply chains and
regulatory policy. Handbook on the Sustainable Supply Chain. Edward Elgar Publishing.
Retrieved April 7th from https://doi.org/10.4337/9781786434272.00034
Dash, R., McMurtey, M., Rebman, C., Kar, U. (2019). Application of Artificial Intelligence
in Automation of Supply Chain Management. Journal of Strategic Innovation and
Sustainability, 14(3).
Dogru, A., & Keskin, B.B. (2020). AI in operations management: applications, challenges
and opportunities. Journal of Data, Information and Management, 2, p.67-74.
Duong Khanh, L.N., & Chong, J. (2020). Supply chain collaboration in the presence of
disruptions: a literature review. International Journal of Production Research, 58(11).
https://doi.org/10.1080/00207543.2020.1712491
Dutta, P., Choi, T-M., Somani, S., & Butala, R. (2020). Blockchain technology in supply
chain operations: Applications, challenges and research opportunities. Transportation
Research Part E: Logistics and Transportation Review, 142.
https://doi.org/10.1016/j.tre.2020.102067
68
Emrouznejad, A., Abbasi, S., & Sıcakyüz, Ç.Supply chain risk management: A content
analysis-based review of existing and emerging topics. (2023). Supply Chain Analytics, 3.
https://doi.org/10.1016/j.sca.2023.100031
Ersoy, Ç., & Camgoz Akdag, H. (2022). Recent Developments in Supply Chain Compliance
in Europe and Its Global Impacts on Businesses. Digitizing Production Systems: Selected
Papers from ISPR2021, October 07-09, 2021 Online, Turkey, 578-584.
Fan, Y., Chen, J., Shirkey, G., John, R., Wu, S.R., Park, H., & Shao, C. (2016). Applications
of structural equation modeling (SEM) in ecological studies: an updated review. Ecological
Processes, 5(19). DOI 10.1186/s13717-016-0063-3
Gammelgaard, B., & Nowkicka, K. (2023). Next generation supply chain management: the
impact of cloud computing. Journal of Enterprise Information Management, 17(1).
https://doi.org/10.1108/JEIM-09-2022-0317
Ganesh. A.D., & Kalpana, P. (2022). Future of artificial intelligence and its influence on
supply chain risk management - A systematic review. Computers & Industrial Engineering,
169. https://doi.org/10.1016/j.cie.2022.108206
Ganga, G.M.D., & Carpinetti, L.C.R. (2011), A fuzzy logic approach to supply chain
performance management. International Journal of Production Economics, 134(1), p.
177-187. https://doi.org/10.1016/j.ijpe.2011.06.011
George, A.S., Hovan George, A.S., & Baskar, T. (2023). Leveraging AI to Revolutionize
Procurement: ChatGPT’s Hidden Potential to Transform the Procurement Iceberg.
International Innovation Journal, 1(6). https://doi.org/10.5281/zenodo.10367172
Ghaffarian, S., Taghikahah, F.R., & Maier, H.R. (2023). Explainable Artificial Intelligence in
disaster risk management: Achievements and prospective futures. International Journal of
Disaster Risk Reduction, 98. https://doi.org/10.1016/j.ijdrr.2023.104123
Ghetto, J. (2021). Analysis and Evaluation of the Impact of Artificial Intelligence on Value
Creation in the Supply Chain. [PhD thesis, University of Gloucestershire]. Analysis and
Evaluation of the Impact of Artificial Intelligence on Value Creation in the Supply Chain -
Research Repository (glos.ac.uk)
Giannakis, M., Spanaki, K., & Dubey, R. (2019). A cloud-based supply chain management
system: effect on supply chain responsiveness. Journal of Enterprise Information
Management, 32(4), p. 585-607. https://doi.org/10.1108/JEIM-05-2018-0106
Glory, E. (2023). Supplier Relationship Management in the Era of Industry 4.0. European
Journal of Supply Chain Management, 1(1), p. 41-51.
69
Govindan, K., & T.C.E., C. (2018). Advances in stochastic programing and robust
optimization for supply chain planning. Computers & Operations Research, 100, p. 262-269.
https://doi.org/10.1016/j.cor.2018.07.027
Greengard, S. (2021). The Internet of Things. MIT Press. The Internet of Things, revised and
updated edition - Samuel Greengard - Google Böcker
Guida, M., Caniato, F., Moretto, A., & Ronchi, S. (2023). The role of artificial intelligence in
the procurement process: State of the art and research agenda. Journal of Purchasing and
Supply Management, 29(2). https://doi.org/10.1016/j.pursup.2023.100823
Habibi, F., Chakrabortty, R.K., & Abbasi, A. (2023). Evaluating supply chain network
resilience considering disruption propagation. Computers & Industrial Engineering, 183.
https://doi.org/10.1016/j.cie.2023.109531
Hangl, J., Behrens Joy, V., & Krause, S. (2022). Barriers, Drivers and Social Considerations
for AI Adoption in Supply Chain Management: A Tertiary Study. Logistics, 6(63).
https://doi.org/10.3390/logistics6030063
He, B., Huang, H., & Yuan, K. (2016). Managing supply disruption through procurement
strategy and price competition. International Journal of Production Research, 54(7), p.
1980-1999. https://doi.org/10.1080/00207543.2015.1074297
Helo, P., & Hao, Y. (2022). Artificial Intelligence in operations management and supply
chain management: an exploratory case study. Production Planning & Control, 33(16),
p.1573-1590. https://doi.org/10.1080/09537287.2021.1882690
Herrera, A., & Janczewski, L. (2015). Cloud supply chain resilience. 2015 Information
Security for South Africa (ISSA), p. 1-9. DOI: 10.1109/ISSA.2015.7335076.
Ho, W., Prasanta, K.D., & Lockström, M. (2011). Strategic sourcing: a combined QFD and
AHP approach in manufacturing. Journal of Supply Chain Management, 16(6).
Hudnurkar, M., Deshpande, S., Rathod, U., & Jakhar, S.K. (2017). Supply Chain Risk
Classification Schemes: A Literature Review. Operations and Supply Chain Management,
10(4), pp. 182-199.
Hu, Y., & Ghadimi, P. (2022). A review of blockchain technology application on supply chain
risk management. IFAC-PapersOnLine, 55(10), pp. 958-963.
https://doi.org/10.1016/j.ifacol.2022.09.472
Ivanov, D. (2021). Supply Chain Risks, Disruptions and Ripple Effect. Introduction to Supply
Chain Resilience Classroom Companion: Business, pp. 1-28.
https://doi.org/10.1007/978-3-030-70490-2_1
70
Ivanov, D., Dolgui, A., Sokolov, B., & Invanova, M. (2017). Literature review on disruption
recovery in the supply chain. International Journal of Production Research, 55(20).
https://doi-org.ezproxy.ub.gu.se/10.1080/00207543.2017.1330572
Jahani, N., Sepehri, A., Vandachali, H.R., & Tirkolaee, E.B. (2021). Application of Industry
4.0 in the Procurement Processes of Supply Chains: A Systematic Literature Review.
Sustainable Global Supply Chain Management, 13(14). https://doi.org/10.3390/su13147520
Jiang, J., Aldewereld, H., Dignum, V., Wang, S., & Baida, Z. (2015). Regulatory compliance
of business processes. AI & Society, 30, p. 393-402.
https://doi.org/10.1007/s00146-014-0536-9
Kallio, H., Pietilä, A-M., Johnson, M., & Kangasniemi, M. (2016). Systematic
methodological review: developing a framework for a qualitative semi-structured interview
guide. Journal of Advanced Nursing, 72(12), p. 2954-2965. https://doi.org/10.1111/jan.13031
Kangra, K., & Singh, J. (2022). Explainable Artificial Intelligence: Concepts and Current
Progression. Explainable Edge AI: A Futuristic Computing Perspective, 1072, p. 1-17.
https://doi.org/10.1007/978-3-031-18292-1_1
Kassa, A., Kitaw, D., Stache, U., Beshah, B., & Degefu, G. (2023). Artificial intelligence
techniques for enhancing supply chain resilience: A systematic literature review, holistic
framework, and future research. Computers & Industrial Engineering, 186.
https://doi.org/10.1016/j.cie.2023.109714
Kashem, M.A., Shamsuddoha, M., Nasir, T., & Chowdhury, A.A. (2023). Supply Chain
Disruption versus Optimization. A Review on Artificial Intelligence and Blockchain.
Knowledge, 3(1), p. 80-86. https://doi.org/10.3390/knowledge3010007
Katok, E., & Tan, L. (2017). Procurement in the Presence of Supply Disruptions. Eindhoven
University of Technology. Assessed 04-11 from DisruptionHoldUp_EK_100918_MS
(utdallas.edu)
Kaur, N., & Kaur, H. (2022). A Multi agent Based Evacuation Planning for Disaster
Management: A Narrative Review. Archives of Computational Methods in Engineering, 29,
pp. 4085-41133.
Khorsan, R., & Crawford, C. (2014). External Validity and Model Validity: A Conceptual
Approach for Systematic Review Methodology. Research Methodology: Choices, Logistics
and Challenges. https://doi.org/10.1155/2014/694804
71
Kochan, C., Nowicki, D.R., Randall, W.S., & Pourreza, S. (2014). Impact of Cloud
Computing on Supply Chain Resilience. (2014). Annual CSCMP Educators Conference.
http://dx.doi.org/10.13140/2.1.4274.7841
Kopetz, H., & Steiner, W. (2022). Internet of Things. Real-Time Systems, p. 325-241.
https://doi.org/10.1007/978-3-031-11992-7_13
Kruhse-Lehtonen, U., & Hofmann, D. (2020). How to Define and Execute Your Data and AI
Strategy. Harvard Data Science Review, 2(3). https://doi.org/10.1162/99608f92.a010feeb
Kulkarni, V., Sunkle, S., Kholkar, D., Roychoudhury, S., Kumar, R., & Raghunandan, M.
(2021). Toward automated regulatory compliance. CSI Transactions on ICT, 9, p. 95-104.
https://doi.org/10.1007/s40012-021-00329-4
Kumar, P., & Kumar, S.V. Enhancing Supplier Selection through Explainable AI: A
Transparent and Interpretable Approach. (2023). International Conference on Research
Methodologies in Knowledge Management, Artificial Intelligence and Telecommunication
Engineering. https://doi.org/10.1109/RMKMATE59243.2023.10369820
Lai, J-Y., Wang, J., & Chiu, Y-H. (2021). Evaluating blockchain technology for reducing
supply chain risks. Information Systems and e-Business Management, 19, pp. 1089-1111.
Lambora, A., Gupta, K., & Chopra, K. (2019). Genetic Algorithm - A Literature Review.
International Conference on Machine Learning, Big Data, Clod and Parallel Computing, p.
380-284. doi: 10.1109/COMITCon.2019.8862255.
Lavastre, O., Gunasekaran, A., & Spalanzani, A. (2014). Effect on firm characteristics,
supplier relationships and techniques used on Supply Chain Risk Management (SCRM): an
empirical investigation on French industrial firms. International Journal of Production
Research, 52(11). https://doi.org/10.1080/00207543.2013.878057
Li, Y., Chen, K., Collignon, S., & Ivanov, D. (2021). Ripple effect in the supply chain
network: Forward and backward disruption propagation, network health and firm
vulnerability. European Journal of Operations Research, 291(3), pp. 1117-1131.
https://doi.org/10.1016/j.ejor.2020.09.053
Lim, Lee, V.-H., Foo, P.-Y., Ooi, K.-B., & Wei–Han Tan, G. (2022). Unfolding the impact of
supply chain quality management practices on sustainability performance: an artificial neural
network approach. Supply Chain Management, 27(5), 611–624.
https://doi.org/10.1108/SCM-03-2021-0129
LoBiondo-Wood, G., & Haber, J. (2014). Nursing Research: Methods and critical appraisal
for evidence-based practice. Elsevier Mosby.
72
Lohmer, J., Budget, N., & Lasch, R. (2020). Analysis of resilience strategies and ripple effect
in blockchain-coordinated supply chains: An agent-based simulation study. International
Journal of Production Economics, 228. https://doi.org/10.1016/j.ijpe.2020.107882
Longo, F., Nicoletti, L., Padovano, A., D’Atri, G., & Forte, M. (2019). Blockchain-enabled
supply chain: An experimental study. Computers & Industrial Engineering, 136, pp. 57-69.
https://doi.org/10.1016/j.cie.2019.07.026
Longo, L., Goebel, R., Lecue, F., Kiseberg, P., & Holzinger, A. (2020). Explainable Artificial
Intelligence: Concepts, Applications, Research Challenges and Visions. Machine Learning
Knowledge Extraction, 12279, p. 1-16. https://doi.org/10.1007/978-3-030-57321-8_1
Lynn, T., Mooney, J.G., Rosati, A., & Cummins, M. (2019). Disruption Finance: FinTech and
Strategy in the 21st century. Palgrave Macmillian e-book. 1007030.pdf (oapen.org)
Maesa, D.D.F., Mori, P., & Ricci, L. (2019). A blockchain based approach for the definition
of auditable Access Control Systems. Computers & Security, 84, p. 93-119.
https://doi.org/10.1016/j.cose.2019.03.016
Maestrini, V., Luzzini, D., Caniato, F., & Ronchi, S. (2018). Effects of monitoring and
incentives on supplier performance: An agency theory perspective. International Journal of
Production Economics, 203, p. 322-332. https://doi.org/10.1016/j.ijpe.2018.07.008
Marley, K.A., Ward, P.T., & Hill, J.A. (2014). Mitigating supply chain disruptions - a normal
accident perspective. Supply Chain Management, 19(2).
Matin, A., Islam, R.M., Wang, X., & Huo, H. (2023). AIoT for sustainable manufacturing:
Overview, challenges and opportunities. Internet of Things, 24.
Matook, S., Lasch, R., & Tamaschke, R. (2009). Supplier development with benchmarking as
a part of a comprehensive supplier risk management framework. International Journal of
Operations & Production Management, 29(3). DOI 10.1108/01443570910938989
Mediavilla, M.A., Dietrich, F., & Palm, D. (2022). Review and analysis of artificial
intelligence methods for demand forecasting in supply chain management. Procedia CIRP,
107, p. 1126-1131.
Meyer, D., & Henke, M. (2023). Developing design principles for the implementation of AI
in PSM: An investigation with expert interviews. Journal of Purchasing and Supply
Management, 29(3). https://doi.org/10.1016/j.pursup.2023.100846
73
Min, H. (2010). Artificial intelligence in supply chain management: theory and applications.
International Journal of Logistics Research and Applications, 13(1), p. 13-29.
10.1080/13675560902736537
Min, H. (2019). Blockchain technology for enhancing supply chain resilience. Business
Horizons, 62(1), pp. 32-45. https://doi.org/10.1016/j.bushor.2018.08.012
Modgil, S., Kumar Singh, R., & Hannibal, C. (2022). Artificial intelligence for supply chain
resilience: learning from Covid-19. The International Journal of Logistics Management, 1(1).
Mo, Y., Harrison, T.P., & Barton, R.R. (2010). Solving stochastic programming models in
supply chain design using sampling heuristics. IMA Journal of Management Mathematics,
22, p. 65-77.
Nair, A., Jayaram, J., & Das, A. (2015). Strategic purchasing participation, supplier selection,
supplier evaluation and purchasing performance. International Journal of Production
Research, 53(20), p. 6263-6278. https://doi.org/10.1080/00207543.2015.1047983
Nimmy, S., Hussain, O.K., Chakrabortty, R.K., Hussain, K.F., Saberi, M. (2022).
Explainability in supply chain operational risk management: A systematic literature review.
Knowledge Based Systems, 225. https://doi.org/10.1016/j.knosys.2021.107587
Nitsche, B, Straube, F., & Wirth, M. (2021). Application areas and antecedents of automation
in logistics and supply chain management: a conceptual framework. Supply Chain Forum: An
International Journal, 22(3), p. 223-239.
https://doiorg.ezproxy.ub.gu.se/10.1080/16258312.2021.1934106
Oliveira, F.A., Nobre, C.N., & Zarate, L.E. (2013). Applying Artificial Neural Networks to
prediction of stock price and improvement of the directional prediction index - Case Study of
PETR4, Petrobras, Brazil. Expert Systems with Applications, 40(18), p. 7596-7606.
https://doi.org/10.1016/j.eswa.2013.06.071
Patel, R., & Davidsson, B. (2019). Forskningsmetodikens Grunder: Att planera, genomföra
och rapportera en undersökning. Studentlitteratur AB.
Peidro, D., Mula, J., Jimenez, M., & Botella, M. (2010). A fuzzy linear programming based
approach for tactical supply chain planning in an uncertainty environment. European Journal
of Operational Research, 205(1), p. 65-80. https://doi.org/10.1016/j.ejor.2009.11.031
Pournader, M., Ghaderi, H., Hassanzadegan, A., & Fahimnia, B. (2021). Artificial
intelligence applications in supply chain management. International Journal of Production
Economics, 241. https://doi.org/10.1016/j.ijpe.2021.108250
74
Praveen, U., Farnaz, G., & Hatim, G. (2019). Inventory management and cost reduction of
supply chain processes using AI based time-series forecasting and ANN modeling. Procedia
Manufacturing, 38, p. 356-263. https://doi.org/10.1016/j.promfg.2020.01.034
Prieto, A., Prieto, B., Ortigosa, E.M., Ros, E., Pelayo, F., Ortega, J., & Rojas, I. Neural
networks: An overview of early research, current frameworks and new challenges.
Neurocomputing, 214, p. 242-268. https://doi.org/10.1016/j.neucom.2016.06.014
Queiroz, M.M., Telles, R., & Bonilla, S.H. (2019). Blockchain and supply chain management
integration: a systematic review of the literature. Supply Chain Management, 25(2), pp.
241-254. https://doi.org/10.1108/SCM-03-2018-0143
Razali, A.R., & Tahir, I.M. (2011). Review of the Literature on Enterprise Risk Management.
Business Management Dynamics, 1(5), p. 08-16.
Renz, S.M., Carrington, J.M., & Badger, T.A. (2018). Two strategies for Qualitative Content
Analysis: An Intramethod Approach to Triangulation. Qualitative Health Research, 28(5).
https://doi.org/10.1177/1049732317753586
Riahi, Y., Saikouk, T., Gunasekaran, A., & Badraoui, I. (2021). Artificial intelligence
applications in supply chain: A descriptive bibliometric analysis and future research
directions. Expert Systems with Applications, 173.
https://doi.org/10.1016/j.eswa.2021.114702
Richey, R.G., Chowdhury, S., David-Sramek, B., Giannakis, M., & Dwivedi, Y.K. (2023).
Artificial intelligence in logistics and supply chain management: A primer and roadmap for
research. Journal of Business Logistics, 44(4), p. 532-549. https://doi.org/10.1111/jbl.12364
Sabia, R. (2019). Artificial Intelligence and Environmental Criminal Compliance.
International Review of Penal Law, p. 179-201.
Sahoo, S.K., Goswami, S.S., & Halder, R. (2024). Supplier Selection in the Age of Industry
4.0: A review on MCDM Applications and Trends. Decision Making Advances, 2(1), p.
32-47. https://doi.org/10.31181/dma21202420
Sallam, K.M., Mohamed, A.W., & Mohamed, M. (2023). Internet of Things (IoT) in Supply
Chain Management: Challenges, Opportunities and Best Practices. SMIJ, 2.
https://doi.org/10.61185/SMIJ
Sarker, S., Engwall, M., Trucco, P., & Feldmann, A. (2016). Internal Visibility of External
Supplier Risks and the Dynamics of Risk Management Silos. IEEE Transactions on
Engineering Management, 63(4), pp. 451-461. doi: 10.1109/TEM.2016.2596144
75
Setak, M., Sharifi, S., & Alimohammadian, A. (2012). Supplier Selection and Order
Allocation Models in Supply Chain Management: A Review. World Applied Sciences
Journal, 18(1), p. 55-72. DOI: 10.5829/idosi.wasj.2012.18.01.3258
Seyedan, & Mafakheri, F. (2020). Predictive big data analytics for supply chain demand
forecasting: methods, applications, and research opportunities. Journal of Big Data, 7(1),
1–22. https://doi.org/10.1186/s40537-020-00329-2
Sharma, R., Shishodia, A., Gunasekaran, A., Min, H., & Munim, Z. H. (2022). The role of
artificial intelligence in supply chain management: mapping the territory. International
Journal of Production Research, 60(24), p. 7527–7550.
https://doi.org/10.1080/00207543.2022.2029611
Sharma, S. (2019). Implementation of ERM Using Artificial Intelligence. ResearchGate.
http://dx.doi.org/10.13140/RG.2.2.36505.47204
Shirvastav, M. (2022). Barriers Related to AI implementation in Supply Chain Management.
Journal of Global Information Management, 30(8), p.1-19. DOI: 10.4018/JGIM.296725.
Shuting, L., & Chen, X. (2019). The role of supplier collaboration and risk management
capabilities in managing product complexity. Operations Management Research, 12, pp.
146-158. https://doi.org/10.1007/s12063-019-00144-w
Silva, N., Ferreira, L.M..D.F., Silva, C., Magalhaes, V., & Neto, P. (2017). Improving Supply
Chain Visibility With Artificial Neural Networks. Procedia Manufacturing, 11, p. 2083-2090.
https://doi.org/10.1016/j.promfg.2017.07.329
Simangunsong, Hendry, L. C., & Stevenson, M. (2012). Supply-chain uncertainty: a review
and theoretical foundation for future research. International Journal of Production Research,
50(16), 4493–4523. https://doi.org/10.1080/00207543.2011.613864
Sodhi, M.S., & Tang, C.S. (2019). Research Opportunities in Supply Chain Transparency.
Production and Operations Management, 28(12). https://doi.org/10.1111/poms.13115
Soori, M., Areezo, B., & Dastres, R. (2023). Artificial neural networks in supply chain
management, a review. Journal of Economy and Technology, 1, p. 179-196.
https://doi.org/10.1016/j.ject.2023.11.002
Sullivan, Y., & Fosso Wamba, S. (2022). Artificial Intelligence, Firm Resilience to Supply
Chain Disruptions and Firm Performance. Hawaii International Conference on System
Sciences, 55.
Sürücü, L., & Maslakci, A. (2020). Validity And Reliability in Quantitative Research. BMIJ,
8(3), pp. 2694-2726. doi: http://dx.doi.org/10.15295/bmij.v8i3.154
76
Tenny, S., Brannan, J.M., & Brannan, G.D. (2017). Qualitative Study. StatPearls Publishing.
Toorajipour, R., Sohrabpour, V., Nazarpour, A., Oghazi, P., & Fischl, M. (2021). Artificial
Intelligence in Supply Chain Management: A systematic review. Journal of Business
Research, 122, p.502-517. https://doi.org/10.1016/j.jbusres.2020.09.009
Van Hoek, R., & Lacity, M. (2023, November). How Global Companies Use AI to Prevent
Supply Chain Disruptions. Harvard Business Review. Assessed from How Global Companies
Use AI to Prevent Supply Chain Disruptions (hbr.org)
Volvo Group. (2023). Geared for Growth.
https://www.volvogroup.com/content/dam/volvo-group/markets/master/investors/reports-and-
presentations/annual-reports/AB-Volvo-Annual-Report-2022.pdf
Wang, B., Lin, Z., Wang, M., Wang, F., Xiangli, P., & Li, Z. (2023). Applying Blockchain
technology to ensure compliance with sustainability standards in the PPE multi-tier supply
chain. International Journal of Production Research, 61(14), p. 4934-4950.
https://doi.org/10.1080/00207543.2022.2025944
Won Bae, J., Shin, K., Lee, H-R., Lee, T., Kim, C.H., Cha, W-C., Kim, G.W., & Moon, I-C.
(2018). Evaluation of Disaster Response System Using Agent-Based Model With Geospatial
and Medical Details. IEEE Transactions on Systems, Man, and Cybernetics Systems, 48(9),
pp. 1454-1469. doi: 10.1109/TSMC.2017.2671340
Xu, F., Uszkoreit, H., Du, Y., Fan, W., Zhao, D., & Zhu, J. (2019). Explainable AI: A Brief
Survey on History, Research Areas, Approaches and Challenges. Natural Language
Processing and Chinese Computing, p. 563-574.
https://doi.org/10.1007/978-3-030-32236-6_51
Yang, Y., Lin, J., Liu, G., & Zhou, L. (2021). The behavioral causes of bullwhip effect in
supply chains: A systematic literature review. International Journal of Production
Economics, 236. https://doi.org/10.1016/j.ijpe.2021.108120
Zhang, J., & Tao, D. (2021). Empowering Things With Intelligence: A Survey of the
Progress, Challenges and Opportunities in Artificial Intelligence of Things. IEEE Internet of
Things Journal, 8(10).
77
Appendices
Appendix A - Interview Questions
1. How are AI technologies today being applied in Volvo AB’s supply chain?
2. How do you mitigate disruptions and risks today?
3. How do you analyze risks?
4. What are your expectations of utilizing AI within supply chain management?
5. When implementing AI did you need to alter the supply chain or did the AI fit into the
supply chain configuration directly?
6. What challenges do you see in implementing AI in supply chain management?
7. Have you considered any specific AI technology or tool to implement as of today that
would fit your supply chain? For example the internet of things?
8. Would you like to implement blockchain? Or is it too challenging?
9. What do you believe is the future for AI in the supply chain? New unexplored
application areas?
78