Received: 30 May 2023 Accepted: 31 May 2023

DOI: 10.1111/jscm.12304

INVITED ARTICLE

Artificial intelligence for supply chain management:

Disruptive innovation or innovative disruption?

Christian Hendriksen

Department of Operations Management,

Copenhagen Business School, Abstract
Frederiksberg, Denmark This article examines the theoretical and practical implications of artificial
intelligence (AI) integration in supply chain management (SCM). AI has
Correspondence
Christian Hendriksen, Assistant Professor, developed dramatically in recent years, embodied by the newest generation of
Department of Operations Management, large language models (LLMs) that exhibit human-like capabilities in various
Copenhagen Business School,
Frederiksberg, Denmark.
domains. However, SCM as a discipline seems unprepared for this potential
Email: [email protected] revolution, as existing perspectives do not capture the potential for disruption
offered by AI tools. Moreover, AI integration in SCM is not only a technical
but also a social process, influenced by human sensemaking and interpretation
of AI systems. This article offers a novel theoretical lens called the AI Integra-
tion (AII) framework, which considers two key dimensions: the level of AI
integration across the supply chain and the role of AI in decision-making. It
also incorporates human meaning-making as an overlaying factor that shapes
AI integration and disruption dynamics. The article demonstrates that differ-
ent ways of integrating AI will lead to different kinds of disruptions, both in
theory and in practice. It also discusses the implications of AI integration for
SCM theorizing and practice, highlighting the need for cross-disciplinary
collaboration and sociotechnical perspectives.

KEYWORDS
artificial intelligence, disruption, GPT-4, metaphorical imagination, sensemaking

INTRODUCTION et al., 2023), text production (Garrido-Merch
an

et al., 2023), understanding other living beings’ point of
Artificial intelligence (AI) has developed dramatically in view (Sileo & Lernould, 2023), and data analysis
a short time, embodied by the newest generation of (Mollick, 2023). The full range of capabilities of these
large language models (LLMs) launched in early 2023. models is not even understood by their creators
For supply chain management (SCM) scholars and (OpenAI, 2023). Still, early evidence suggests that AI
practitioners alike, this potential revolution necessitates models are exceptionally powerful, surpassing humans
re-evaluating the existing theoretical framework and in some advanced cognitive tests (Bubeck et al., 2023).
practical applications. AI models are approaching or As SCM scholars, we must ask ourselves: If these sys-
even transcending human-like capabilities in areas such tems are as capable as they seem, how will this change
as apparent reasoning (Espejel et al., 2023; Liu supply chains and SCM research?

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
© 2023 The Author. Journal of Supply Chain Management published by Wiley Periodicals LLC.

J. Supply Chain Manag. 2023;59:65–76. wileyonlinelibrary.com/journal/jscm 65

66 JOURNAL OF SUPPLY CHAIN MANAGEMENT

However, SCM as a discipline seems unprepared at a chain where virtually all functions are carried out by
theoretical level for this revolution. Some researchers autonomous AI agents. Crucially, I demonstrate that AI
have recently explored the possibilities of digitalization integration in supply chains is a technical, operational,
and AI in making supply chains more efficient (Perano and social process, and future SCM theorizing on this
et al., 2023), while others have studied the potential of AI topic must take seriously this complexity.
systems in mitigating disruptions caused by the COVID-
19 pandemic (Nayal et al., 2022). While these perspectives
are valuable, they do not capture the potential for disrup- THE EVOLUTION OF AI AND I TS
tion offered by AI tools. There are two primary reasons IM P ACT O N SCM
for this: First, newer and more capable AI systems are
personalized chatbots. Unlike previous AI technological Previous iterations of different AI systems have revolved
systems, this generation of AI does not need a large infra- around machine learning (ML) and big data analytics to
structure to work; anyone with an internet connection achieve efficiency gains or quality improvement (Kamble
can start using these AI tools right away, which elimi- et al., 2021). This was the predominant perspective before
nates a significant barrier to AI adoption. Second, the December 2022, focusing on how ML infrastructure
capability of the most advanced systems, especially could be deployed to analyze very large amounts of data
GPT-4, is at a level where it can assist a professional to provide humans with conclusions and results (Dubey
human being in a large array of complex knowledge et al., 2020). ML approaches have required extensive
intensive or creative tasks (Bubeck et al., 2023; architecture and AI expertise, given that ML solutions
Mollick, 2023). This capability also entails challenges as would require local setup in organizations and special-
we risk relying too much on AI tools, put undue trust in ized experts that could train the models to give useful
AI judgments, or allow AI-induced bias to cause problems output (Kinra et al., 2020).
for minorities and vulnerable groups. In other words, we It is no longer an exaggeration to say that this per-
are now living in a reality where every person on the spective was obsolete on the morning of December
planet—and thus, every SCM professional—has personal 1, 2022. The day before, OpenAI launched a conversa-
access to the most powerful AI system developed in tional chatbot called ChatGPT, which could produce
human history. We are now experiencing the beginning moderately complex text relating to a wide variety of
stages of a true fourth industrial revolution (Gates, 2023). topics. ChatGPT was powered by the GPT architecture, a
In this article, I offer a novel theoretical lens that relatively new form of AI that did not analyze input (like
serves as a vehicle for understanding the range of possi- existing ML approaches) but predicted output based on
ble disruptions and supply chain impacts brought about user input. This approach, known as generative AI, did
by increasingly capable AI systems. While the concept not replace ML systems, but it suddenly allowed everyone
“disruption” is sometimes used without clear definitions to open ChatGPT and chat with the AI about relatively
in SCM (Ketchen & Craighead, 2021; van den Adel complex topics. Moreover, unlike ML systems where only
et al., 2022), I understand disruptions in SCM to be AI experts understood the systems and could operate or
events that alter the flow of goods or services, requiring train them, ChatGPT could generate output for any user
supply chain adaptation and innovation, and affecting regardless of training or background. This is the first rea-
social and environmental outcomes. I couple the theoreti- son business approaches to AI changed so rapidly:
cal lens with an interpretivist perspective (Darby ChatGPT demonstrated that AI systems could be made
et al., 2019) on how AI is understood by actors in the sup- available and approachable to anyone, not just a select
ply chain through a sensemaking perspective few AI engineers.
(Weick, 1995). This novel perspective, which I call the AI In March 2023, OpenAI revealed their new iteration
Integration (AII) framework, is designed to provide a of the GPT architecture called GPT-4. This model exhibits
novel perspective on AI integration in SCM. In this arti- reasoning capabilities that outrank average humans on
cle, “integration” refers to the increasingly extensive use virtually every evaluation parameter and surpasses the
of AI systems to carry out work tasks throughout the sup- best 5% of humans on most evaluation parameters
ply chain. The purpose of the framework in combination (Bubeck et al., 2023; OpenAI, 2023). When GPT-4 was
with the interpretivist lens is to demonstrate that, given the ability to write, interpret, and execute code, it
depending on the trajectory and form of AI integration became capable of running fairly complex data analysis
happening in a given supply chain, different sets of possi- and visualizing (Mollick, 2023). When used simply as a
ble disruptions emerge. For example, a supply chain chatbot, the standard model allowed any person to pro-
where AI is only sparingly used entails a completely dif- duce fairly complex material, with some even showing
ferent set of potential disruptions compared with a supply that GPT-4 could replace humans doing research
ARTIFICIAL INTELLIGENCE FOR SUPPLY CHAIN MANAGEMENT 67

(Hitch, 2023). GPT-4 is the most powerful generative AI it changes the rules of the game for our theoretical
model at the time of writing, but other models are being models, assumptions, and frameworks. The reason for
developed in parallel by other companies, such as Goo- this is that the theories we use in SCM are predicated on
gle. This competitive race to produce ever-more powerful the idea that humans and organizations are the building
AI systems that are widely accessible will ensure that blocks of supply chains, and supply chains emerge as
increasingly capable AI systems will be in the hands of phenomena from these interactions (Carter et al., 2015;
professionals sooner rather than later. Choi et al., 2001; Wieland, 2021). Yet, as AI becomes
In SCM, there are already indications of how this will increasingly capable of not only generating text, analysis,
change the landscape. For example, early indications sug- and media but also making decisions and adjudicating
gest that GPT-4 can carry out the analytical basis for sup- choices, it calls into question to what extent our existing
plier choice, such as evaluating a supplier profile based assumptions still hold.
on a given set of parameters and instructions (provided To exemplify the broader theoretical implications,
by humans) (Bonde, 2023). It can also be integrated into consider two theories used in SCM: transaction cost eco-
sales, demand forecasting, production and logistics opti- nomics (TCE) (Hobbs, 1996; Ketokivi & Mahoney, 2020;
mization, and customer sales, to name a few Tate et al., 2014) and complex adaptive systems (CAS)
(Panigrahi, 2023). However, the broader implications go (Choi et al., 2001; Nair & Reed-Tsochas, 2019). These two
beyond these items: How does it change a procurement theories are good exemplars because they entail different
manager’s job when they have an expert-level AI at hand assumptions and are used by scholars to explain different
to assist in task solutions, workflow structuring, brain- SCM phenomena. In TCE, the core focus is how to struc-
storming, writing, and analytics? And how does a supply ture the governance of a supply chain (or any relation-
chain manager approach the well-known challenge of ship or transaction) in such a way that it is the most
managing a complex supply chain beyond their horizon efficient, given objectives (Ketokivi & Mahoney, 2020,
(Carter et al., 2015) when they have the capability of an p. 1012; Williamson, 1979). Two foundational assump-
AI readily available to help? I suspect these are questions tions in this perspective are that actors are boundedly
managers may ask themselves sooner rather than later. rational and opportunistic if given a chance, and because
However, AI does not only pose a possibility for firms of this, actors are constrained in their ability to engage in
and individual professionals to improve productivity and efficient transactions depending on the context. But now,
quality. AI integration entails many pitfalls and dangers, imagine a situation where two actors have extensive AI
such as the potential of overreliance on AI systems for assistance. This immediately changes the idea of bounded
carrying out tasks, delegating authority to AI systems rationality, or at least relaxes this constraint, because any
when it is inappropriate, undue faith in AI evaluation, or given actor has extensive capability to analyze and act
unethical usage of AI systems with regard to employees upon information.
or customers (Boiko et al., 2023; Panigrahi, 2023; Similarly, sufficiently advanced AI systems may guide
Shevlane et al., 2023). In this sense, the possibility for dis- users towards long-term benefits rather than short-term
ruption in supply chains is multidimensional. One dimen- gains. Long-term utility maximization for all transaction
sion of disruption relates to the promissory nature of participants is a better outcome for a time-independent
generative AI, given its approachability and personalized AI. Perhaps most insidiously, we can ask ourselves: In a
manifestation, where everyone can use it for their day- situation where it is no longer humans but rather autono-
to-day tasks. Another dimension of disruption can emerge mous AI systems that negotiate with each other, how
from the unintended consequences of any combination of does TCE even work? What can we assume about AI
pitfalls because these can be exacerbated throughout the behavior in such a situation?
supply chain. For example, if several companies in a CAS theories have been integral to SCM theorizing
chain use AI to produce analyses that are wrong or do not since the early 2000s, with the underlying concept being
have the capability to use AI responsibly, what happens that supply chains emerge from the interaction of inter-
in the entire chain when an AI-based system breaks nal mechanisms within the system and the external envi-
down? Could there be an “AI-bullwhip effect”? ronment, leading to coevolution in a chaotic and
nonlinear fashion (Choi et al., 2001; Nair & Reed-
Tsochas, 2019, p. 81). This paradigm may be fundamen-
A I A D O PT I O N A S A DI S R U P TI O N OF tally altered in a context where advanced generative AI
THEORY becomes a ubiquitous tool available to all professionals
across different firms.
As AI systems become more widespread and widely The system’s unpredictability may amplify substan-
adopted at both the individual and organizational levels, tially as AI-driven workflows swiftly transform across the
68 JOURNAL OF SUPPLY CHAIN MANAGEMENT

supply chain. Concurrently, the “environment”— developed as a theoretical model designed to capture the
encompassing markets, consumer behaviors, regulatory complexities and nuances of AI integration in supply
landscapes, and even competing supply chains— chains and to contextualize a certain disruptive potential
increasingly integrates AI technologies. These factors, depending more easily on the specifics of the AI integra-
changing and adapting quickly in less predictable ways, tion in a given supply chain. Additionally, while AI is
set the stage for supply chains and their sociopolitical used by individuals or firms that are part of larger supply
contexts to coevolve at an unprecedented pace. For chains, the emergent (Choi et al., 2001; Sawyer, 2004)
instance, AI systems could react instantaneously to fluc- effects of AI manifest at the supply chain level as disrup-
tuating demand patterns identified by AI-enabled market tions that are larger than the sum of their parts. Accord-
analysis tools or adjust supply chain processes in ingly, the AII framework operates at the supply chain
response to regulatory changes. This AI-driven dyna- level because the emergent effects stemming from indi-
mism in the environment introduces a new layer of com- vidualized technologies can be more disruptive. The
plexity to supply chains. More fundamentally, deeply framework is based on two key dimensions: the level of
integrated or autonomous AI systems could challenge AI integration across the supply chain and the role of AI
CAS’s traditional explanations for supply chain emer- in decision-making. By considering these two dimen-
gence, as the role of “actor” shifts from humans to AI sys- sions, the AII provides a comprehensive perspective on
tems. This may redefine measures of information and AI integration in SCM, capturing the potential disrup-
entropy in CAS, as generative AI could pioneer novel tions and transformations that can occur due to AI
communication and computation modes that exceed tra- adoption.
ditional metrics. Furthermore, AI could enable new However, we know from recent SCM contributions
forms of adaptation and learning, surpassing the bounds that human interpretations of systems and technologies
of human agency and rationality as understood in tradi- matter for how SCM practice emerges and practices are
tional CAS theories (Yu et al., 2023). Consequently, the shaped (Darby et al., 2019, 2022; Wieland, 2021). Thus, I
introduction of generative AI into SCM necessitates a propose that in addition to considering the level of AI
critical reassessment of CAS theories concerning system integration and the role of AI in decision-making,
emergence and complexity. scholars should pay attention to the way supply chain
This is a partial outline of the potential theoretical professionals engage in the sensemaking of AI systems
import of the fourth industrial revolution. Yet, it shows and how they ascribe meaning and identity to them
that even with some very simple examples, the core ideas (Scott, 2014, pp. 55–63; Weick, 1995; Weick et al., 2005).
of these theories should be considered in this new light. As humans mentally categorize AI systems and make
New SCM research in the “age of AI” (Gates, 2023) sense of them, humans maintain agency over how,
should work on clarifying what our theories can and can- whether, and why AI integration in a supply chain mani-
not handle, how we should develop or revise our assump- fests. The framework consists of the two axes
tions about the world, and how we theoretically explain mentioned—yielding a 2  2 model—but with human
supply chain mechanisms and dynamics emerging from meaning-making overlaid and given as an interpretation
increasing integration of AI. of how AI-induced disruptions can play out in a given
supply chain.
The following paragraphs help explain the two axes
T H E AI I F R A M E W O R K I N SC M and how the 2  2 model yields ways of understanding
different ways AI can be integrated into supply chains.
As the possible disruptions to both theory and First, metaphors are attached to each quadrant to assist
practice stemming from AI integration become clearer, in imagining the kind of supply chains and work situa-
SCM needs a way to categorize and understand AI tions that may arise from each combination (Stephens
integration across supply chains. In the simplest terms, et al., 2021; Tsoukas, 1993). Then, the explanation shifts
different ways of integrating AI will lead to different to explain how socially constructed sensemaking pro-
kinds of disruptions, both in theory and in practice. cesses give rise to different understandings of what AI
For example, light and surface-level AI integration across can and should do in a supply chain. Taking all this
a supply chain is a different challenge than a supply together, the potential for disruptions in each quadrant
chain that operates almost entirely on autonomous AI is discussed, including how sensemaking can yield
agents. agency to humans in steering this and what this means
Accordingly, SCM scholars can understand the trajec- for the trajectory of AI integration. Finally, it explains
tory of AI integration and its associated potential for dis- how these disruptions can change SCM scholarship and
ruptions through the lens of the AII framework. AII is practice.
ARTIFICIAL INTELLIGENCE FOR SUPPLY CHAIN MANAGEMENT 69

Dimensions of the AII framework humans and AI systems, the accountability and control
mechanisms that need to be in place, and the ethical
The AII is based on two key dimensions: the level of AI implications of delegating decision-making power to AI
integration across the supply chain and the role of AI in systems.
decision-making. The level of AI integration across the As noted earlier, the model operates at the level of
supply chain refers to the extent to which AI is integrated the supply chain. As individual firms implement AI into
into the various activities and processes of the supply their operations, AI dynamics at the supply chain level
chain. At one end of the spectrum, AI integration is low, becomes an emergent phenomenon (Choi et al., 2001;
with few actors in the supply chain adopting AI or only Sawyer, 2004). Firms will most likely begin with partial
using AI for specific tasks. This could be likened to the integration with AI in an assistive role, but depending on
early stages of AI adoption, where AI is used as a tool for the characteristics of the supply chain in question, firms
specific tasks such as demand forecasting or inventory may then either move towards other quadrants or main-
management but in a limited capacity. In this stage, AI tain limited AI integration. This depends on the specifics
acts as an assisting system, enhancing human capabilities of the supply chain in question and the operational situa-
but not replacing them, and never making autonomous tion in the firms involved. The important thing to note
decisions or executing without human direction. here is that firm-level AI implementation strategies lead
At the peak of AI integration, numerous supply chain to AI integration dynamics across the supply chain.
actors incorporate AI deeply into their operations, span- These dimensions can be seen in Figure 1. In quad-
ning tasks from demand forecasting to delivery and wider rant one, the metaphor for understanding (Stephens
operations management. High integration is not merely et al., 2021) is Human Sherlock, Robot Watson, referring
about extensive infrastructure within a single firm but to the classic fictional universe of the detective Sherlock
refers to supply chain actors leveraging AI tools like Holmes. Here, AI is partially integrated and serves in an
ChatGPT across all job functions. While the tools might assistive role. That means humans are in the lead, and AI
be individualized, their collective usage across the supply is used sparingly in select use cases across the supply
chain leads to an emergent, robust AI-driven supply chain. In this way, the Human Sherlock has full owner-
chain. The versatility of these tools outmatches earlier ship of the tasks, sets directions, and makes decisions,
ML infrastructure solutions, allowing for effective use while Robot Watson does a lot of the legwork, pulls
within various tasks across firms. Despite varied integra- together material, and follows instructions as best as pos-
tion strategies across companies, the emergent nature of sible. Picture a warehouse where employees are selecting
AI usage contributes to unique supply chain dynamics. items for order fulfillment. AI assists by scanning inven-
The second axis is the role of AI in decision-making. tory data and providing the optimal route for picking
It refers to the extent to which AI systems are given items, thus aiding human workers without making the
autonomy and decision-making power. At one end of the ultimate decisions. Human Sherlock directs operations,
spectrum, AI plays an assistive role, providing data analy- while Robot Watson expedites tasks and offers advice
sis, finding information, and providing 1:1 assistance for based on data. However, like Watson, AI can also chal-
employees. In this role, AI is primarily a tool that lenge Human Sherlock’s ideas when they are unreason-
enhances human capabilities, providing support for able or go against best available knowledge, and it can
decision-making but not making decisions autono- make Human Sherlock aware that there are perspectives
mously. This can be likened to the role of a trusted advi- Sherlock may be missing. Thus, while Sherlock is still in
sor or assistant, who provides valuable insights and control and solves the case, Watson serves as a valuable
recommendations but does not have the authority to asset not only in terms of menial legwork but also in an
make decisions. intellectual capacity.
AI takes on an autonomous role at the other end of As we increase integration, we move towards
the spectrum, analyzing and executing independently. In quadrant two. Here, the metaphor is that of a Robot
this role, AI systems are given autonomy and decision- Cartographer that lays out the entire map and relates it
making power, potentially making decisions that signifi- to the territory (Fabbe-Costes et al., 2020). This means
cantly impact the supply chain. This can be likened to the AI is used in most or all advanced and general
the role of a manager or director, who provides insights functions, conducting data analysis, planning, and
and recommendations and has the authority to make organizing information inside and outside the
decisions and implement changes. This shift from an organizational boundary. Consequently, integration in
assistive to an autonomous role represents a significant this quadrant will be more focused and organized by
change in supply chain dynamics. Furthermore, it raises the firm in question. However, the AI only does this to
crucial questions about the balance of power between assist humans in making final decisions. An example
70 JOURNAL OF SUPPLY CHAIN MANAGEMENT

F I G U R E 1 Artificial
Intelligence (AI) Integration
framework for supply chains.

could be a complex multi-echelon supply chain scenario supply chain transforms into an AI-led ecosystem. Envi-
where AI is employed to forecast demand, identify sion a supply chain where AI is used to control all
supply chain risks, monitor supplier performance, and aspects—from demand forecasting, inventory manage-
suggest optimal distribution strategies. While AI lays ment, order fulfillment to logistics and customer service.
out the entire map of data and strategies, the human Each AI system communicates and cooperates with others,
“captain” uses this information to make final decisions. making autonomous decisions based on set instructions
In this quadrant, humans universally take on supervisory and reacting to changes in real time. This ecosystem learns
roles in the supply chain, making decisions and and adapts, forming a complex, interconnected network
organizing ideas and strategies based on ubiquitous AI that operates with minimal human intervention.
assistance in all functions. Moving from the theoretical metaphors that demon-
Even with partial integration, we can give AI systems strate various stages of AI integration, it is essential to
authority to make decisions, leading us to quadrant three. understand that these are not just abstract notions.
Here, the metaphor is Chess Grandmaster, where the AI Instead, these metaphors represent the reality of how AI
system is used in specific tasks but is also allowed to interacts with and influences the supply chain process,
make decisions itself. Much like a Chess Grandmaster is with significant implications for SCM. However, the true
expert at a very narrow set of choices relating to the game impact of AI depends on more than just its integration or
of chess, AI systems in these supply chains are used autonomy. It is also contingent on the perception of the
deliberately on specific tasks. In this instance, AI might individuals interacting with these systems—the human
autonomously analyze different vendors based on price, agents in our supply chain. This brings us to an impor-
delivery times, reliability, and past performance. It then tant aspect of AI integration: the human sensemaking of
selects the best option without direct human interven- AI systems and the potential for disruption. The human
tion. However, given that AI systems in these situations understanding of AI, shaped by their experiences and
interface with human-led systems, humans still have perspectives, significantly influences the interpretation
final authority to override AI choices if it conflicts with and acceptance of these systems within the supply chain.
other parts of the workstreams or production flows.
Finally, in quadrant four, supply chains may be orga-
nized so that AI is both fully integrated into all functions Human sensemaking of AI systems and the
and given the authority to make decisions autonomously. potential for disruption
The metaphor here is that of an AI Ecosystem. When all
actors in a supply chain have fully integrated AI capabili- The integration of AI into supply chains is not merely a
ties in all their main tasks and give the AIs authority to technical or operational endeavor; it deeply intersects
decide and act (given some predefined instructions), the with the realm of the social. To grasp the intricacies of
ARTIFICIAL INTELLIGENCE FOR SUPPLY CHAIN MANAGEMENT 71

this integration, we must consider the pivotal role of think about how to integrate it into existing structures.
human actors who actively interact with these AI systems For example, suppose humans in a supply chain under-
daily. They interpret the potential and limitations of stand AI systems to be a threat (due to security issues,
these technologies, manage their implementation, and displacement of human jobs, or vulnerability due to tech-
guide their utilization. Their perceptions and beliefs nological dependency). In that case, integration will be
about AI, shaped by their individual experiences, profes- done with that in mind—i.e., AI integration on noncriti-
sional backgrounds, and their specific roles within the cal issues, separation from human tasks, firewalling from
supply chain, significantly influence the degree of AI core processes, and more. Alternatively, a supply chain
integration and its role in supply chain operations. These that understands AI as a competitive capability will aim
human actors determine whether AI is deployed mini- to enhance all supply chain actors’ ability to integrate AI
mally or widely across various supply chain activities, to ensure the supply chain is competitive.
whether it plays an assistive or autonomous role, and the Depending on the sensemaking processes that play
potential disruptions that could arise from its use. out, the potential for disruptions in each quadrant will
Understanding the phenomenon of human sensemak- pan out differently. In quadrant one, humans who under-
ing, the process by which people give meaning to their stand Robot Watson as a threat will be more hesitant to
collective experiences, is paramount in this context. Their rely on AI output or use AI systems in critical work pro-
individual interpretations, borne from their unique cesses. This means disruptions in the supply chain as a
beliefs, experiences, and professional viewpoints, are inte- whole are less likely given that human actors are careful
gral in shaping the trajectory of AI adoption and its ulti- not to allow AI systems to threaten core functions. How-
mate utility in the supply chain context. For instance, a ever, if Robot Watson is seen as a useful partner, disrup-
procurement manager with a positive outlook towards AI tive potential increases but is contingent on the supply
might see it as a strategic tool to streamline procurement chain practices that are already in place due to the role as
decisions, while a logistics manager who harbors skepti- a partner. When integration increases and we move to a
cism about AI’s reliability might view it as a potential risk. Robot Cartographer, human actors may emphasize the
These divergent interpretations can significantly shape risk of the Cartographer producing wrong maps—i.e.,
how AI is deployed and managed within the same supply producing wrong inferences—if they think of the AI as a
chain, thereby determining the potential benefits, chal- tool. This also means that the type of disruption changes,
lenges, and disruptions that could emerge. Recognizing from a potential systemic shock to a supply chain to a
this social aspect of AI integration, alongside the critical limited problem that is local to a few supply chain actors.
technical considerations, is vital for effectively managing Thus, the set of possible disruptions emerging from a
AI technologies within supply chains. This dual under- given type of integration depends critically on how
standing equips organizations with the necessary insights human actors make sense of these systems.
to harness the potential of AI, minimize its risks, and The general picture that emerges is one where
thereby create more resilient and efficient supply chains. human sensemaking is evocative of the agency human
While each quadrant represents its own set of possible actors have in directing AI integration. Different meta-
disruptions both in practice and in theory, it is important phors and interpretations of AI systems lead to different
to emphasize that the approach to AI integration in any ways of thinking about how AI systems should be
supply chain is contingent on humans’ interpretation of integrated, what risk assessments and risk management
these systems (Berger & Luckmann, 1967; Darby strategies supply chain actors should implement, and
et al., 2019; Wieland, 2021). AI integration is not just a what type of AI integration a company should aim for,
technical process—it is also a social process. The basic with a given velocity of change. All this happens in a
premise here is that humans, when faced with a new sit- given supply chain context (Bille & Hendriksen, 2022)
uation, phenomenon, or technology, interpret this novel that shapes the specific mechanisms. Based on the AII
event by relating it to existing experience (Bendoly, 2016; combined with the sensemaking perspective, it is
Weick et al., 2005). This process is idiosyncratic across important to underline that future SCM research on AI
organizations and supply chains, resulting in divergences integration must recognize the social component of AI
from supply chain to supply chain in terms of how AI is integration. Approaching AI integration with a purely
understood and interpreted. In Table 1, I summarize a technical perspective risks ignoring the considerable
few main types of interpretation, but not exhaustive, as importance of social interactions that give rise to
social sensemaking processes are complex and defy neat differential AI implementation strategies. With this, I
categorization. will now turn to a forward-looking discussion of future
Each way of understanding and interpreting what AI theorizing in SCM and the practical implications arising
is and its role leads to vastly different ways that humans from these ideas.
72 JOURNAL OF SUPPLY CHAIN MANAGEMENT

TABLE 1 Sensemaking of AI and implications for SCM theories and supply chain dynamics.

Sensemaking
of AI Impact on theoretical assumptions in SCM Changes in supply chain dynamics
AI as a tool AI is viewed as an instrument that enhances human AI leads to automation of routine tasks, reducing
capabilities. Theories in SCM would need to manual errors and increasing operational efficiency.
incorporate the idea of AI as a powerful extension It can also lead to job displacement due to
of human decision-making, reducing transaction automation.
costs, and improving efficiency.
AI as an actor AI is perceived as an autonomous entity capable of AI takes over certain decision-making tasks, leading to
decision-making. Theories must grapple with the significant changes in roles and responsibilities
concept of AI as an active participant in SCM, within the supply chain. It also presents potential
challenging traditional definitions of an “actor” in challenges in accountability and control.
the supply chain.
AI as a partner AI is seen as a collaborator, working alongside human AI creates a human–AI collaboration in decision-
actors. Theories need to consider the synergistic making, leading to enhanced decision quality.
relationship between humans and AI, focusing on However, it could also lead to overreliance on AI or
collaborative decision-making and shared issues with trust and acceptance among human
responsibilities. workers.
AI as a threat AI is perceived as a potential risk, displacing human AI could lead to job displacement, new vulnerabilities
workers or introducing new vulnerabilities. like AI system failures or cyber-attacks, and
Theories would need to factor in the potential potential social and ethical issues.
downsides and risks associated with AI integration.
AI as a AI is seen as a strategic asset that can provide AI enhances the competitive positioning of the
competitive competitive advantage. SCM theories would need to organization, driving innovation and improving
capability integrate the concept of AI as a source of customer service. However, it could also lead to
differentiation and competitive edge. increased competition and the risk of falling behind
if not adopted effectively.
AI as a factor AI is considered as a new category of production factor, AI changes the cost structure and efficiency of supply
of production akin to labor or capital. Theories would need to re- chain operations. It could also lead to new business
evaluate traditional factors of production to include models and value creation strategies.
AI.
AI as relief Theories must incorporate AI’s role in alleviating AI handles routine tasks, increasing efficiency. This
mundane tasks and enabling strategic focus for leads to shifts in human roles towards more strategic
humans. tasks, possibly introducing a skills gap.

Abbreviations: AI, artificial intelligence; SCM, supply chain management.

I M P L I C A T I O N S FO R T H E O R I Z I N G theories towards acknowledging AI as more than just a

I N S C M AN D T H E N A T U R E O F AI passive tool in human decision-making. Recognizing AI
DISRUPTION S I N P R ACT I CE as active participants in supply chain interactions sig-
nifies a shift in perspective that acknowledges the agency
I will now delve into the broader implications and and autonomy of AI systems. These systems are no lon-
forward-looking perspectives. Drawing on the insights ger mere extensions of human ability but possess their
and evidence that have been discussed so far, I aim to own capacities to influence and shape supply chain
highlight the transformative role of AI and the accompa- dynamics. From this perspective, it does not matter
nying challenges and opportunities. whether AI systems are sentient or self-conscious.
Instead, it is important to recognize that AI systems by
virtue of their capabilities can impact supply chains in
AI as active participants in SCM similar ways to humans.
These active AI systems engage in intricate interac-
The exploration of AI integration in SCM, underpinned tions with humans that go beyond one-way manipula-
by the AII framework and our understanding of the ways tion. They coevolve and exert mutual influences,
humans interact with AI, reveals a need to reorient SCM suggesting that theories of AI in SCM might need to
ARTIFICIAL INTELLIGENCE FOR SUPPLY CHAIN MANAGEMENT 73

center around understanding human–AI relationships AI disruptions and the role of social
and collaboration. Much like how previous SCM processes
literature has deciphered human collaboration across
supply chains, future theories could aim to demystify The disruption that AI brings to SCM is not just techno-
AI–human relations in this context (Blome et al., 2014; logically determined but also shaped by social processes,
Danese et al., 2020; Frohlich & Westbrook, 2001). adding another layer of complexity to theorizing in SCM.
Depending on the trajectory of AI development and inte- This warrants the evolution of sociotechnical theories
gration, this could be an important topic for future SCM that can capture the interplay between the technical
theorists. aspects of AI integration and the social reactions and
responses of human actors. These theories might need to
examine the role of organizational culture, norms, and
New constructs and philosophical values in shaping AI adoption and use, along with the
perspectives in SCM theorizing sensemaking processes guiding AI interpretation. The
impact of these interpretations on AI integration and dis-
Incorporating AI as active participants necessitates ruption dynamics should also be taken into account
that our theories accommodate new constructs, such as (Bille & Hendriksen, 2022; Darby et al., 2019; Touboulic
the idea of “AI autonomy”—which encapsulates the et al., 2020; Wieland, 2021).
extent to which AI systems can operate independently, Moreover, AI integration in SCM offers a unique
making decisions without human intervention. Addition- opportunity for cross-disciplinary collaboration. SCM
ally, there is the potential construct of “AI–human researchers can partner with experts from diverse fields
collaboration,” representing the synergistic interaction like computer science, engineering, sociology, psychol-
between humans and AI and its impact on supply chain ogy, and more to develop a more nuanced understanding
performance. It is crucial that theorists also face the of AI and its implications for supply chains (Kamble
possible conflicts that might emerge from AI integration, et al., 2021; Trautrims et al., 2016). This could lead to the-
like power struggles between human and AI actors, oretical insights that break through the typical con-
ethical considerations of AI decision-making, and the straints of SCM theorizing, integrating concepts from
control mechanisms required to regulate AI behavior. human–computer interaction, information systems, data
This is particularly important given the potential for analytics, and behavioral sciences (Bogers et al., 2017).
AI-induced disruptions stemming from irresponsible use
of AI.
To fully comprehend the integration of AI in SCM, Practical implications and cross-
researchers may need to pull from multiple philosophical disciplinary collaboration for AI in SCM
perspectives. The infusion of AI into supply chains is a
complex, unpredictable process involving many variables When considering practical implications for everyday
and possible outcomes. Thus, the integration of positivist, supply chain practices, AI integration presents both
interpretivist, and critical realist perspectives can aid in opportunities and challenges. Organizations must pay
providing a more comprehensive understanding. A posi- attention to not only the technical aspects of AI adoption
tivist approach can help to identify general patterns and but also the human elements, such as managing human
relationships associated with AI integration, whereas an perceptions and reactions towards AI, fostering an AI-
interpretivist perspective can illuminate the nuanced aware culture, and providing training for effective AI–
ways humans interpret and interact with AI. A critical human collaboration. Ethical and accountability issues
realist perspective can recognize both the existence of associated with delegating decision-making power to AI
objective realities in AI-driven SCM operations and the must also be addressed. Control mechanisms to prevent
influence of subjective interpretations on these realities misuse and monitor AI behavior, as well as potential job
while shifting focus to causal mechanisms underpinning displacement and other social issues resulting from AI
disruptive effects (Beach & Pedersen, 2019; Illari & integration, need to be considered.
Williamson, 2012). Ironically, AI systems like GPT-4 AI disruptions can vary across supply chains, requir-
allow SCM researchers to consider and evaluate different ing a systemic approach to manage them. This includes
philosophical perspectives much easier. This is because creating robust systems for monitoring AI performance,
the model can provide a perspective or argument from establishing measures for disruption mitigation, and
another philosophical point of view, allowing researchers developing contingency plans. Coordination and collabo-
to evaluate different perspectives much faster and easier ration with other supply chain actors—aligning AI strate-
than before. gies and developing shared best practices for AI
74 JOURNAL OF SUPPLY CHAIN MANAGEMENT

integration and management—are integral to this research and debate on this important topic and inspire
approach. However, at a more fundamental, supply chain new ways of thinking about and managing supply chains
managers must consider whether they have capabilities in the AI era.
to integrate AI systems in a meaningful way
(Barney, 1991; Teece et al., 1997). This entails operational A C KN O WL ED G EME N T S
considerations and responsible management practices. A I am grateful to Tingting Yan and Wendy Tate for their
situation where firms integrate AI into core functions excellent feedback and valuable input.
without considering how to do it responsibly constitutes
a major risk for disruptive effects. Consequently, supply ORCID
chain managers must carefully consider how AI integra- Christian Hendriksen https://orcid.org/0000-0003-0987-
tion can be done in a responsible manner. 329X
As we progress deeper into this era of AI-driven sup-
ply chains, we must balance the use of AI for efficiency RE FER EN CES
with maintaining transparency, fairness, and account- Barney, J. (1991). Firm resources and sustained competitive
ability. This requires continuous revisiting, refining, and advantage. Journal of Management, 17, 99–120. http://jom.
reimagining our understanding and management of AI sagepub.com/content/17/1/99.short. https://doi.org/10.1177/
in SCM. As we forge ahead, we carry the agency over 014920639101700108
Beach, D., & Pedersen, R. B. (2019). Process-tracing methods:
how AI can and will be integrated into supply chains.
Foundations and guidelines (2nd ed.). University of Michigan
But such an integration demands a solid theoretical Press. https://doi.org/10.3998/mpub.10072208
grounding upon which we can construct our Bendoly, E. (2016). Fit, bias, and enacted sensemaking in data
understanding—a foundation that can help us shape the visualization: Frameworks for continuous development in
future of SCM in the AI era. operations and supply chain management analytics. Journal
of Business Logistics, 37(1), 6–17. https://doi.org/10.1111/jbl.
12113
Berger, P. L., & Luckmann, T. (1967). The social construction of real-
C O N C L U S IO N
ity: A treatise in the sociology of knowledge. Doubleday.
Bille, A., & Hendriksen, C. (2022). Let us get contextual: Critical
In this article, I have argued that AI integration in SCM realist case studies in supply chain management. Supply Chain
is a potential revolution that challenges existing theories Management: An International Journal, 28, 724–737.
and practices. I have proposed the AII framework as a https://doi.org/10.1108/SCM-03-2022-0119
novel theoretical lens to understand how AI can be inte- Blome, C., Paulraj, A., & Schuetz, K. (2014). Supply chain collabora-
grated into supply chains and the associated potential for tion and sustainability: A profile deviation analysis. Interna-
disruptions. I have also highlighted the role of human tional Journal of Operations & Production Management, 34(5),
639–663. https://doi.org/10.1108/IJOPM-11-2012-0515
sensemaking processes in shaping the interpretation and
Bogers, M., Zobel, A.-K., Afuah, A., Almirall, E., Brunswicker, S.,
adoption of AI systems in SCM. I have suggested that
Dahlander, L., Frederiksen, L., Gawer, A., Gruber, M., &
SCM scholars and practitioners need to revisit their Haefliger, S. (2017). The open innovation research landscape:
assumptions and frameworks to account for the agency Established perspectives and emerging themes across different
and autonomy of AI systems, the interplay between tech- levels of analysis. Industry and Innovation, 24(1), 8–40.
nical and social aspects of AI integration, and the ethical https://doi.org/10.1080/13662716.2016.1240068
and accountability issues that arise from delegating Boiko, D. A., MacKnight, R., & Gomes, G. (2023). Emergent autono-
decision-making power to AI systems. mous scientific research capabilities of large language models.
Bonde, F. (2023). Navigating the technical landscape—Large
The AII framework is not meant to be a definitive or
language models such as “GPT-4” in business, trade ….
comprehensive model of AI integration in SCM but Medium. https://medium.com/@frederikbonde/navigating-
rather a starting point for further exploration and dia- the-technical-landscape-large-language-models-such-as-gpt-4-
logue. As AI systems evolve and become more capable, in-business-trade-f0862f8edf5d
the AII must be updated and refined to capture new Bubeck, S., Chandrasekaran, V., Eldan, R., Gehrke, J., Horvitz, E.,
developments and applications. Moreover, the AII can be Kamar, E., Lee, P., Lee, Y. T., Li, Y., Lundberg, S., Nori, H.,
complemented by other theoretical perspectives that can Palangi, H., Ribeiro, M. T., & Zhang, Y. (2023). Sparks of
offer deeper insights into specific aspects of AI integra- artificial general intelligence: Early experiments with GPT-4
(arXiv:2303.12712). arXiv. http://arxiv.org/abs/2303.12712
tion, such as institutional theory, stakeholder theory, or
Carter, C. R., Rogers, D. S., & Choi, T. Y. (2015). Toward the theory
resource-based view. Future research can also empirically of the supply chain. Journal of Supply Chain Management,
test and validate the AII using a more complete AI- 51(2), 89–97. https://doi.org/10.1111/jscm.12073
enabled toolbox of various theoretical and methodologi- Choi, T. Y., Dooley, K. J., & Rungtusanatham, M. (2001). Supply
cal approaches. This article will stimulate further networks and complex adaptive systems: Control versus
ARTIFICIAL INTELLIGENCE FOR SUPPLY CHAIN MANAGEMENT 75

emergence. Journal of Operations Management, 19, 351–366. normal states. Journal of Supply Chain Management, 57(1),
https://doi.org/10.1016/S0272-6963(00)00068-1 50–57. https://doi.org/10.1111/jscm.12251
Danese, P., Molinaro, M., & Romano, P. (2020). Investigating fit in Ketokivi, M., & Mahoney, J. T. (2020). Transaction cost economics
supply chain integration: A systematic literature review on as a theory of supply chain efficiency. Production and Opera-
context, practices, performance links. Journal of Purchasing tions Management, 29(4), 1011–1031. https://doi.org/10.1111/
and Supply Management, 26(5), 100634. https://doi.org/10. poms.13148
1016/j.pursup.2020.100634 Kinra, A., Hald, K. S., Mukkamala, R. R., & Vatrapu, R. (2020).
Darby, J. L., Fugate, B. S., & Murray, J. B. (2019). Interpretive An unstructured big data approach for country logistics
research: A complementary approach to seeking knowledge in performance assessment in global supply chains. International
supply chain management. International Journal of Logistics Journal of Operations & Production Management, 40(4), 439–
Management, 30, 395–413. https://doi.org/10.1108/IJLM-07- 458. https://doi.org/10.1108/IJOPM-07-2019-0544
2018-0187 Liu, H., Ning, R., Teng, Z., Liu, J., Zhou, Q., & Zhang, Y.
Darby, J. L., Fugate, B. S., & Murray, J. B. (2022). The role of small (2023). Evaluating the logical reasoning ability of ChatGPT
and medium enterprise and family business distinctions in and GPT-4 (arXiv:2304.03439). arXiv. http://arxiv.org/abs/
decision-making: Insights from the farm echelon. Decision Sci- 2304.03439
ences, 53(3), 578–597. https://doi.org/10.1111/deci.12538 Mollick, E. (2023). It is starting to get strange. https://www.
Dubey, R., Gunasekaran, A., Childe, S. J., Bryde, D. J., oneusefulthing.org/p/it-is-starting-to-get-strange
Giannakis, M., Foropon, C., Roubaud, D., & Hazen, B. T. Nair, A., & Reed-Tsochas, F. (2019). Revisiting the complex adap-
(2020). Big data analytics and artificial intelligence pathway to tive systems paradigm: Leading perspectives for researching
operational performance under the effects of entrepreneurial operations and supply chain management issues. Journal of
orientation and environmental dynamism: A study of Operations Management, 65(2), 80–92. https://doi.org/10.1002/
manufacturing organisations. International Journal of Produc- joom.1022
tion Economics, 226. Scopus, 107599. https://doi.org/10.1016/j. Nayal, K., Raut, R., Priyadarshinee, P., Narkhede, B. E.,
ijpe.2019.107599 Kazancoglu, Y., & Narwane, V. (2022). Exploring the role of
Espejel, J. L., Ettifouri, E. H., Alassan, M. S. Y., Chouham, E. M., & artificial intelligence in managing agricultural supply chain
Dahhane, W. (2023). GPT-3.5 vs GPT-4: Evaluating ChatGPT’s risk to counter the impacts of the COVID-19 pandemic. The
reasoning performance in zero-shot learning. Arxiv. International Journal of Logistics Management, 33(3), 744–772.
Fabbe-Costes, N., Lechaptois, L., & Spring, M. (2020). “The map is https://doi.org/10.1108/IJLM-12-2020-0493
not the territory”: A boundary objects perspective on supply OpenAI. (2023). GPT-4 technical report (arXiv:2303.08774). arXiv.
chain mapping. International Journal of Operations & Produc- http://arxiv.org/abs/2303.08774
tion Management, 40(9), 1475–1497. https://doi.org/10.1108/ Panigrahi, S. (2023). How GPT-4 can change the game in supply
IJOPM-12-2019-0828 chain management. https://www.linkedin.com/pulse/how-gpt-
Frohlich, M. T., & Westbrook, R. (2001). Arcs of integration: An 4-can-change-game-supply-chain-management-suraj-panigrahi/
international study of supply chain strategies. Journal of Oper- Perano, M., Cammarano, A., Varriale, V., Del Regno, C.,
ations Management, 19(2), 185–200. https://doi.org/10.1016/ Michelino, F., & Caputo, M. (2023). Embracing supply chain
S0272-6963(00)00055-3 digitalization and unphysicalization to enhance supply chain
Garrido-Merch
an, E. C., Arroyo-Barrigüete, J. L., & Gozalo- performance: A conceptual framework. International Journal
Brizuela, R. (2023). Simulating H.P. Lovecraft horror literature of Physical Distribution and Logistics Management. https://doi.
with the ChatGPT large language model (arXiv:2305.03429). org/10.1108/IJPDLM-06-2022-0201
arXiv. http://arxiv.org/abs/2305.03429 Sawyer, R. K. (2004). The mechanisms of emergence. Philosophy of
Gates, B. (2023). The age of AI has begun. Gatesnotes.Com. https:// the Social Sciences, 34, 260–282. https://doi.org/10.1177/
www.gatesnotes.com/The-Age-of-AI-Has-Begun 0048393103262553
Hitch, D. (2023). Artificial intelligence (AI) augmented qualitative Scott, W. R. (2014). Institutions and organizations: Ideas, interests,
analysis: The way of the future? (SSRN Scholarly Paper and identities (4th ed.). Sage Publications.
No. 4451740). https://papers.ssrn.com/abstract=4451740 Shevlane, T., Farquhar, S., Garfinkel, B., Phuong, M.,
Hobbs, J. E. (1996). A transaction cost approach to supply chain Whittlestone, J., Leung, J., Kokotajlo, D., Marchal, N.,
management. Supply Chain Management, 1(2), 15–27. https:// Anderljung, M., Kolt, N., Ho, L., Siddarth, D., Avin, S.,
doi.org/10.1108/13598549610155260 Hawkins, W., Kim, B., Gabriel, I., Bolina, V., Clark, J.,
Illari, P., & Williamson, J. (2012). What is a mechanism? Thinking Bengio, Y., … Dafoe, A. (2023). Model evaluation for extreme
about mechanisms across the sciences. European Journal for risks (arXiv:2305.15324). arXiv. http://arxiv.org/abs/2305.
Philosophy of Science, 2(1), 119–135. https://doi.org/10.1007/ 15324
s13194-011-0038-2 Sileo, D., & Lernould, A. (2023). MindGames: Targeting theory of
Kamble, S. S., Gunasekaran, A., Kumar, V., Belhadi, A., & mind in large language models with dynamic epistemic modal
Foropon, C. (2021). A machine learning based approach for logic (arXiv:2305.03353). arXiv. https://doi.org/10.48550/arXiv.
predicting blockchain adoption in supply chain. Technological 2305.03353
Forecasting and Social Change, 163. Scopus, 120465. https:// Stephens, V., Matthews, L., Cornelissen, J. P., & Rowlands, H.
doi.org/10.1016/j.techfore.2020.120465 (2021). Building novel supply chain theory using “metaphori-
Ketchen, D. J., & Craighead, C. W. (2021). Toward a theory of cal imagination”. Journal of Supply Chain Management, 58(1),
supply chain entrepreneurial embeddedness in disrupted and 124–139. https://doi.org/10.1111/jscm.12257
76 JOURNAL OF SUPPLY CHAIN MANAGEMENT

Tate, W. L., Ellram, L. M., & Dooley, K. J. (2014). The impact of Weick, K. E. (1995). Sensemaking in organizations. Sage
transaction costs and institutional pressure on supplier Publications.
environmental practices. International Journal of Physical Weick, K. E., Sutcliffe, K. M., & Obstfeld, D. (2005). Organizing and
Distribution and Logistics Management, 44(5), 353–372. the process of sensemaking. Organization Science, 16, 409–421.
https://doi.org/10.1108/IJPDLM-12-2012-0356 https://doi.org/10.1287/orsc.1050.0133
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities Wieland, A. (2021). Dancing the supply chain: Toward transforma-
and strategic management. Strategic Management Journal, tive supply chain management. Journal of Supply Chain
18(7), 509–533. https://doi.org/10.1002/(SICI)1097-0266 Management, 57(1), 58–73. https://doi.org/10.1111/jscm.12248
(199708)18:7<509::AID-SMJ882>3.0.CO;2-Z Williamson, O. E. (1979). Transaction-cost economics: The gover-
Touboulic, A., McCarthy, L., & Matthews, L. (2020). Re-imagining nance of contractual relations. The Journal of Law and
supply chain challenges through critical engaged research. Economics, 22(2), 233–261. https://doi.org/10.1086/466942
Journal of Supply Chain Management, 56(2), 36–51. Yu, Y., Lakemond, N., & Holmberg, G. (2023). AI in the context of
https://doi.org/10.1111/jscm.12226 complex intelligent systems: Engineering management conse-
Trautrims, A., Defee, C., & Farris, T. (2016). Preparing business stu- quences. IEEE Transactions on Engineering Management,
dents for workplace reality—Using global virtual teams in 1–14. https://doi.org/10.1109/TEM.2023.3268340
logistics and SCM education. International Journal of Logistics
Management, 27(3), 886–907. https://doi.org/10.1108/IJLM-01-
2015-0003
Tsoukas, H. (1993). Analogical reasoning and knowledge genera- How to cite this article: Hendriksen, C. (2023).
tion in organization theory. Organization Studies, 14(3), Artificial intelligence for supply chain
323–346. https://doi.org/10.1177/017084069301400301 management: Disruptive innovation or innovative
van den Adel, M. J., de Vries, T. A., & van Donk, D. P. (2022). Resil- disruption? Journal of Supply Chain Management,
ience in interorganizational networks: Dealing with day-to-day
59(3), 65–76. https://doi.org/10.1111/jscm.12304
disruptions in critical infrastructures. Supply Chain Manage-
ment, 27(7), 64–78. https://doi.org/10.1108/SCM-03-2021-
0136/FULL/HTML