ARTIFICIAL INTELLIGENCE IN STRATEGIZING:

PROSPECTS AND CHALLENGES

Georg von Krogh, Shiko Ben-Menahem, Yash Raj Shrestha

ETH Zurich

Cite as: von Krogh, Ben-Menahem & Shrestha (Forthcoming) Artificial Intelligence in

Strategizing: Prospects and Challenges

Acknowledgements: This work has been funded by Swiss National Science Foundation Grant

100013_197763

Abstract

Recent developments in the theory and research on artificial intelligence hold great
promises as well as challenges for the strategist’s core activities and conduct of
strategic processes. These promises and challenges requires the strategy field to both
reevaluate some of the principal assumptions and implications of strategizing. In this
chapter we take stock of research on AI applied to strategizing, and illustrate what we
believe are key questions for future research on the strategy-AI nexus. We discuss the
potential of AI in two stages in the strategy process: strategic analysis and -
formulation, as well as strategy implementation. Our aim is to engage strategy scholars
in advancing AI-related research on strategizing.

Keywords: Strategy, Artificial Intelligence, Strategy-Ai nexus, future of strategy, AI

augmentation
Introduction

The past decade has seen momentous scientific advances in the field of artificial intelligence

(AI), “a system’s ability to interpret external data correctly, to learn from such data, and to use

those learnings to achieve specific goals and tasks through flexible adaptation” (Haenlein &

Kaplan, 2019). Powered by a surge in data generation and computing power, algorithms are

becoming increasingly capable of digitally performing operations that efficiently and effectively

emulate human learning, profound judgment and decision-making across a wide range of

application areas. Such applications include optimization of internal business operations, product

design, the capture of scarce external knowledge, and the screening and recruiting of talent

(Shrestha, Ben-Menahem, & von Krogh, 2019; Shrestha, Krishna, & von Krogh, 2021).

With these developments, management scholars have shown a growing interest in AI’s

potential to support or transform organizations (von Krogh, 2018). Yet, while the application of

AI algorithms is by now widely studied across functional domains including engineering, human

resource management, operations management, information systems, economics, finance,

accounting, and marketing, much less attention has been devoted to understanding how the onset

of AI changes the nature of strategizing—that is, the activities of people involved in the strategy

process (Jarzabkowski, Balogun, & Seidl, 2007; Whittington, 2014).

In this chapter, we take stock of the promise and perils of AI in shaping the future of

strategizing. We discuss applications known to us during the key stages in the strategy process

and identify opportunities for future research on the strategy-AI nexus. Recognizing the historic

diversity of definitions and approaches to strategic management (Nag, Hembrick, & Chen, 2007;

Randall, Dent, 2019), we take a broad perspective and define the strategic management process
as “the full set of commitments, decisions, and actions required for a firm to achieve strategic

competitiveness and earn above-average returns” (Hitt, Ireland, & Hoskisson, 2015).

To structure our discussion, we discern two critical stages (Randall, Dent, 2019) in the

strategy process: strategic analysis (the evaluation of an organization’s external and internal

context), as well as strategy formulation and implementation (the generation, evaluation, and

selection of strategic options and choices). Our goal is not an all-inclusive treatment of how AI

may be used in each of these stages, but to illustrate what we believe are central challenges and

opportunities that AI poses for strategic leaders in the foreseeable future.

The chapter is organized as follows. The next section offers a very brief introduction to

AI with a focus on machine learning (ML) algorithms. In the third section, we discuss

applications of AI in strategic analysis, strategy formulation, and strategy implementation. The

fourth section discusses risks and challenges to strategizing posed by AI, and the final section

outlines research imperatives for strategy scholars. Table 1 displays selected key insights from

this chapter.

<Insert Table 1 here>

What Is Artificial Intelligence?

John McCarthy first coined the term “artificial intelligence” in 1956 as “the science and

engineering of making intelligent machines” (Hamet & Tremblay, 2017). During the 1950s,

early advances in the study of AI coincided with progress in the areas of computer science and

cognitive sciences. Pioneers such as Herbert Simon, Alan Newell, and Marvin Minsky pursued a

strong ambition formulated by Alan Turing, to create a machine capable of emulating intelligent

behavior on par with human beings. Turing’s ambition is still preserved in the form of the Turing

test where programmers strive to create AI that passes this “imitation game.”
 The AI boom of the 1950s was followed by periods of reduced AI research and overall

pessimism related to technology itself, popularly known as “AI winters” (McCorduck, 2004).

Various problems in technical designs, the lack of processing power, and scarcity of high-quality

data to train AI algorithms created severe obstacles in developing a useful form of AI (see von

Krogh & Roos, 1995; McCarthy, 1981). During the past two decades, AI research and its

applications in ML algorithms and deep learning have regained much scientific and practical

attention (Franz, Shrestha, & Paudel, 2020). Four developments contributed to this renaissance:

data generation and systematic storage and delivery, algorithm development, processing power,

and open-source code (von Krogh, 2018).

Machine learning can be understood as a computer program’s ability to learn from

experience regarding a task in a way that it improves its performance with experience. Based on

the data and the goal of use, ML algorithms can be broadly categorized into three classes: (a)

supervised learning, (b) unsupervised learning, and (c) reinforcement learning (see Shrestha et

al., 2020). Each of these ML classes holds different promise for strategists. Supervised ML

algorithms excel at prediction and forecasting by learning from the outcomes of past experiences,

similar to experiential learning in humans. Applications include demand forecasting and

forecasting investment returns. In contrast, unsupervised ML algorithms are useful for

identifying robust patterns in data, by performing tasks such as clustering and dimension

reduction. For example, by identifying clusters of companies following similar strategies and

competing in similar markets, unsupervised ML can aid in strategic group analysis. Finally,

reinforcement learning can be useful for strategists in running simulations and experimenting

with innovation ideas or evaluating the outcome of a given strategic process. Such an application
of reinforcement learning can be useful for strategists in evaluating various alternatives and their

counterfactuals before taking a particular action, leading to organizational learning.

AI and Strategic Analysis

A starting point for exploring the role of AI in the strategy process is its potential for augmenting

and transforming strategic analysis. While information systems have long assisted managers in

gathering and organizing information for environmental scanning and decision-making purposes

(Robinson, Ahmad, & Simmons, 2020), AI algorithms may have the potential for dramatically

expanding the scale, scope, and speed of analysis of an organization’s external and internal

environment through its capacity for interacting with the environment and generating data.

Specifically, AI can assist with (a) foraging information from internal and external data sources

on a discrete or continuous basis, (b) analyzing and interpreting such data through pattern

identification, and (c) assisting executives in decision making through predictive analysis

(Ferràs-Hernández, Tarrats-Pons, & Arimany-Serrat, 2017).

As a starting point for mapping how AI may support strategic analysis, we first discuss its

applications for analyzing external (i.e., macro- and micro-environmental) factors. Next, we

consider AI’s role in competitor analysis, and subsequently in the internal analysis of human,

financial, and auxiliary resources.

External analysis

AI applications have found their way into areas of strategic analysis across political, economic,

social, technological, environmental, and legal factors.

Political scientists have developed various AI-based text analysis tools useful in inferring

the “political climate” across regions. These systems analyze diverse data sources including

news outlets, legislative debates, and online political discourse. In particular, ML algorithms
have proven effective in analyzing unstructured social media data (Zafarani, Abbasi, & Liu,

2014). Using such data, ML is increasingly useful for predicting outcomes of elections (e.g.,

Coletto et al., 2015), policy changes (e.g., Chan & Zhong, 2018), and political bias and conflicts

(Biessmann et al., 2016). Early identification and prediction of political trends and developments

with AI can help strategists understand how government and government policy may influence

their organization. Doing so may provide an important source of competitive advantage.

Machine learning can also be applied to accurately and reliably measure economic trends,

such as economic growth, the onset of economic recessions (e.g., Wu et al., 2020), increasing

poverty (Kshirsagar et al., 2017), and bankruptcies (Cielen, Peeters, & Vanhoof, 2004). The

adoption of ML tools by the financial services industry is an important precursor to applications

by strategists. For example, institutional investors use ML to predict stock returns and thereby

make better investment decisions (Avramov, Cheng, & Metzker, 2019) and estimate systemic

financial risks (e.g., Kou et al., 2019). Economists increasingly apply ML methods to

unconventional data sources and questions, such as the use of satellite image data to predict

spatial distribution of economic well-being across African countries (Jean et al., 2016).

Strategists often need to include social factors in their analysis, such as changes in

demographics and societal values. Some of these changes may be slow-moving (e.g., national

culture), while others are subject to sudden jolts and rapid turns (e.g., fashion) that may pose

significant opportunities or threats to the business. Scholars have proposed techniques such as

automated text analysis to “map the contours of cultural fields, classify cultural elements and

trace the evolution of culture over time” (Bail, 2014). New algorithms also enable systematic

measurement of culture and modeling of its evolution within organizations and social groups
(Doyle et al., 2017). Such developments offer new opportunities for measuring difficult-to-

capture concepts such as culture in strategy research.

Researchers tend to give technological factors an important role in most strategic

analysis, whether for manufacturer or user firms. AI can assist companies in monitoring

technological developments and anticipate any relevant technological changes. For example, ML

is increasingly used to identify patterns of technology development by sifting through massive

amounts of patent or publication data (e.g., Lee et al., 2018). These uses of ML allow firms to

construct “knowledge profiles” of their industry and major competitors (Suominen, Toivanen,

Seppänen, 2017).

With the growing awareness of climate change and the attention given to CSR (corporate

sustainability responsibility), ecological and environmental factors rapidly gain importance in

strategic analysis. Research on the use of AI in such analysis is still in its infancy, but there are

promising examples, such as research into whether hydrogen electrical vehicles may become the

dominant means of transportation for consumers (Ranaei et al., 2016). Such applications are

important for commercial investors, as well as for policy makers deciding on how and where to

develop transportation infrastructures.

Traditionally, firms tend to invest significantly in legal consultants who often manually

collect and analyze a large body of new laws and regulations that may alter the conditions for

doing business across regions and legal regimes. AI applications may support firm internal

collection and processing of such legal data, potentially reducing overall legal expenditures. For

example, Yousfi-Monod, Farzindar, and Lapalme (2010) developed an ML algorithm that sifts

through thousands of pages of legal documents, and provides useful summaries relevant to a
firm’s strategy. AI is also increasingly being used for automating financial compliance

monitoring and regulation (e.g., Treleaven & Batrinca, 2017).

Competitive analysis

A prime application of AI in strategic analysis concerns competitor analysis. As firms gather

massive amounts of data on competitors and their industries, automated analysis can help cluster

data, identify patterns in data, and even predict competitors’ strategic moves.

Often before a strategic analysis may proceed, strategists need to identify relevant

competitors, yet doing so remains one of the strategist’s most challenging tasks (Haefliger, Jäger,

& von Krogh, 2010; Hamel & Prahalad, 1994). New approaches, such as Pant and Sheng’s

(2015) analysis of web content to identify how firms’ service and product offerings “link” or

converge, offer new ways for identifying emerging competitors and predicting performance. In a

similar vein, Shrestha, He, Puranam & von Krogh (forthcoming) propose that ML algorithms can

be used to identify strategic groups consisting of firms that follow similar strategies, and also

predict movements of firms between these groups.

Additionally, such tools offer a more dynamic approach to performing competitive

analysis. Indeed, scholars have warned that assumptions about competitors need to be frequently

re-analyzed and re-examined so as to judge the continuous validity of a firm’s strategy (Ireland,

Hitt, & Sirmon, 2003). The advantage of automated competitor analysis over traditional (manual

and discrete) approaches is considerable: once plugged into market and competitor data

generation machines (e.g., web crawlers, online sales), algorithms can perform competitor

analysis on a continuous basis and help strategists dynamically gauge the validity of the firm’s

current strategy.

Internal analysis
AI may also augment strategic analysis of internal organizational factors, including human,

financial, and auxiliary resources, such as supply chains or customer relationships.

Starting with human resources, AI applications can offer automated pattern recognition in

such data and assist managers in identifying employee performance, predicting career

trajectories, and revealing patterns of compensation and inequality, among others (see

Strohmeier & Piazza, 2013 for an overview). Careful assessment of skills, education, talent, and

demographics enables the strategist to more adequately assess the extent to which a potential

strategy can be implemented.

Firms also possess swaths of accurate, deep, and labeled accounting and financial data.

Strategists tend to draw heavily on balance-sheet information, for example, when conducting

analysis. ML algorithms may continuously or discretely cluster such data, performing analysis of

patterns. In so doing algorithms may boost the interpretation of what causes certain financial

resource conditions (e.g., fluctuations in liquidity, exchange rate premiums at corporate

treasuries) relevant to the type and timing of the firm’s strategic commitments (Fethi &

Pasiouras, 2010). AI also can also be put to use in financial analyses less essential to the strategy

process, such as tools used to monitor transactions performed throughout the firm. Applications

can detect unusual amounts or frequency of transactions, and rapidly identify fraud. Using these

tools, AI can aid in managing financial risk by augmenting compliance and risk management

functions with improved data-driven insights, drawing on rapid and automated data analysis and

reduced administrative loads, thereby saving finance and accounting employees’ time for more

creative activities.

Machine learning algorithms are increasingly providing efficient analysis on a range of

other resources relevant to the strategy process, such as demand forecasting, production
planning, resource allocation, and logistics. For example, ML can augment production planning

decisions by automating the process of searching for potential suppliers by mining data from

online catalogues and other repositories (Nissen & Sengupta, 2006), providing predictions on

performance of prospective suppliers (Humphreys, McIvor, & Huang, 2002), and even

estimating valuation and evaluating online bids (Cheung et al., 2004).

Another auxiliary resource specifically relevant for ML-supported analysis concerns

customer relationships. Customers are increasingly involved in a variety of product- and service-

related discussions, consume digital content, and share knowledge among themselves on a

diverse set of platforms such as YouTube, Facebook, Wikipedia, and TikTok (Hollebeek,

Conduit, & Brodie, 2016). Observation of customer behavior by their digital traces can aid the

strategist in better understanding trends and seasonality in customer needs related to their service

and product offerings (Heinonen, 2011). Such insights can then be used to predict customer

preferences, and eventually lead to analysis relevant for product innovation in the firm (Gomez-

Uribe and Hunt, 2015).

AI in Strategy Formulation and Implementation

With AI’s rapid emergence to general-purpose technology status, numerous organizations are

placing it at the center of their high-level strategy. Google CEO Sundar Pichai emphasized the

company’s strategy as “AI first” in his I/0 2017 keynote, emphasizing a new focus on ML and

other AI technologies to increase organizational efficiency and improve customer experience. In

the subsequent years, Google and its industry partners invested heavily in seemingly more

“intelligent” products such as Google Photos, which uses AI to detect people, places and objects

in images; RankBrain, which increases search speed; and Google Assistant as a personal virtual

assistant--all decisions that markedly increased the overall valuation of the company.
 AI is also reshaping the activities that constitute the formulation and implementation

stages of the strategy process. In their pursuit of profitable growth, firms formulate business

development and financial objectives and corresponding implementation plans at various levels

of granularity. Regardless of whether such a process occurs regularly and formally or

sporadically and informally, formulating and implementing a strategy is a lengthy and complex

process that involves an extensive amount of varied data and draws on managerial analysis and

decision making at various levels of the organization. The cognitive limitations of strategists in

processing data for the decision-making and problem-solving activities underlying strategy

formulation and implementation are therefore well documented in the literature (March &

Simon, 1958; Simon, 1947). The potential of AI to enhance human strategy formulation and

implementation is a contentious subject that much resembles discussions following the

advancement of previous generations of decision support systems since the 1950s (e.g., Yoo &

Digman, 1987).

On one hand, researchers seem to agree that current applications of AI—combined with

the availability of big data—exceed the capabilities of previous systems. Indeed, successful

strategy formulation relies on in-depth knowledge of the organization, its environment, and the

risk-reward profile of strategic options. In a similar vein, successful strategy implementation

relies on a strategist’s ability to effectively monitor operations and evaluate the performance of

strategic actions. The potential of AI-powered systems to automate continuous data analysis,

create new knowledge about strategic opportunities, and identify patterns to predict the payoff of

choices thus stands to enhance the quality of the strategist’s decisions where they concern

formulation and implementation tasks requiring primarily explicit knowledge (Agrawal,

Chatterjee, & Novotný, 2018; Keding, 2020).

 On the other hand, concerns about the limitations of AI-based systems in the strategy

process abound. For example, researchers have argued that while AI performs well in stable,

context-specific settings, it underperforms in situations involving creativity, novelty, and

uncertainty (Brynjolfsson & Mitchell, 2017; Jarrahi, 2018; von Krogh, 2018), and can even

increase complexity in organizational decision making (Lawrence, 1991). Rather than replacing

automated formulation and implementation, AI can thus better be considered as a useful

“assistant” that augments strategists’ capacity to make judgments about external threats and

opportunities, internal strengths and weaknesses, and strategic issues (Huang et al., 2019;

Reeves, Levin, & Ueda, 2016; Shrestha et al., 2019).

AI and corporate strategy

To illustrate how AI can be applied in specific areas of strategy formulation and implementation,

consider its role in the analysis of typical corporate-level strategy questions concerning a firm’s

business portfolio and decisions regarding growth and diversification. AI can be used to model

the performance of a business portfolio; evaluate the fit, risk, and performance of a firm’s set of

businesses; and analyze and propose potential synergies among them. For example, based on an

analysis of market and product data across business units, AI can assist strategists in addressing

portfolio risks, such as when fluctuations in demand are strongly correlated across the portfolio

or when various businesses are exposed to similar currency risks. Davenport refers to such

“sparring partner” activities by AI as “cognitive insight” (Davenport & Ronanki, 2018).

AI tools are also increasingly shaping the way in which companies identify and screen

merger and acquisition opportunities, as well as manage the deal cycle and post-merger

integration (PMI) stage. Whereas traditional mergers and acquisitions (M&A) require a host of

resource-intensive analytic activities (due diligence, market analysis, valuation, pricing) and is
thus limited to screening a small subset of potential targets, AI-based analytics allow companies

to automatically and continuously screen a large pool of opportunities, as they arise. Natural

language processing (NLP) methods such as automatic document summarization and topic

modeling can be used to design a filter such that attractive cases can be preselected for further

human consideration. The strength of AI in these scenarios derives from the technology’s ability

to integrate various data sources--such as patent databases, company financial records, historic

data on M&A deals, social media (e.g., LinkedIn data to identify rare skill sets), news media,

conference call transcript filings, and discussion fora--and dynamically adapt screening

parameters to changing environmental conditions. Emerging financial-technology solution

providers, such as San Francisco-based Refinitiv, use AI tools to develop predictive quantitative

models on M&A targets that help decision makers to more accurately estimate synergies and

assess deal value.1

Moreover, given the critical role of efficiency and speed in decision making around

M&A opportunities (Baum & Wally, 2003) and the post-merger integration stage (Homburg,

Bucerius, 2006), AI can have important performance implications by automating acquisition

process steps, and shortening critical activities in the M&A cycle. These activities may also be

performed at lower cost. For example, with respect to the pre-acquisition stage, recent years have

seen consequential advancements in the use of AI in transactional law, where AI enables legal

professionals to automate the labor-intensive, M&A due-diligence process. Among the many

opportunities for AI, automation can allow target firms to more efficiently collect and classify

relevant documents and ensure that they meet disclosure requirements.

1
See [Link]
 Post-acquisition, AI and social analytics tools can support leaders in organizational

integration by automating the process of mapping human resources and re-deploying staff into

new organizational structures, taking into consideration complex combinations of skills,

remuneration, and job titles (see also “AI and Strategic Analysis” section). AI can further support

and accelerate task integration and identify opportunities for cost reduction by combining large

swaths of unstructured data from separate company systems, such as those found in customer

databases and contract repositories.

In a similar way, AI may find applications in cooperative strategies such as the

identification and selection of strategic alliance partners. For example, NLP of annual reports

may help strategists identify relevant indicators such as a potential partner’s business focus or

strategic positioning (e.g., Menon, Choi, & Tabakovic, 2018).

Strategic innovation, entrepreneurship, and renewal

New insights are currently emerging on how the diffusion of AI technology across firms might

create opportunities and challenges for new venture creation and strategic entrepreneurship.

Cockburn, Henderson, and Stern (2018) argue that AI, and in particular deep learning, serves as

a general-purpose technology (GPT) and a new “method of inventing” with the potential to re-

shape the nature of the innovation process and the organization of R&D. They argue that

advancements in AI, much like the rise of other GPT such as the microprocessor, bring about a

reinforcing cycle between innovation on the level of AI and its application areas.

Chalmers, MacKenzie, and Carter (2020) identify three ways in which AI can enhance

information search and idea generation activities that lie at the basis of new venture opportunity

identification and exploitation. First, the strength of deep-learning algorithms in identifying

structures in high-dimensional data allows corporate and start-up entrepreneurs to search for and
experiment with opportunities that were previously unobservable (e.g., Kucukkeles, Ben-

Menahem, & von Krogh, 2019). Second, AI can support new venture activity through the

identification and exploitation of consumer needs. Third, AI-based simulations and experiments

promise investors and business leaders the ability to test innovations and new venture ideas. For

example, simulations using real-world dynamic data can emulate the conditions new ventures

may encounter in the future. Such tools, when fully developed, may not only benefit investment

decisions, but could one day be used to design innovations and make choices on critical product

features such as pricing or promotions. Using AI-based simulations and experiments,

entrepreneurs could thus test their ideas and predict how customers would react to particular

product features and determine changes in product design and pricing (Chalmers et al., 2020).

AI is also making inroads into augmenting and replacing human activities related to the

exploitation of new venture ideas via the emerging conversational systems (e.g., [Link]) and

robo sales advisors (Chalmers et al., 2020). The extent to which such solutions prove valuable in

the long run remains an open question. Future research opportunities will arise from questions

surrounding the implications of organizational de-skilling and loss of customer relationships as a

result of robo-advisors, as well as questions concerning customer acceptance of interaction with

AI-based systems. From a strategy perspective, important questions arise as to how AI-based

technology changes strategic entrepreneurship when it comes to growing new ventures and

managing the risk inherent in entrepreneurial activity.

Strategy control

Strategy control concerns the efforts of strategists to ensure that the implementation of strategic

plans unfolds as intended and to measure progress on strategic goals. Research on the application

of AI to strategy control has been limited thus far. Nonetheless, several promising areas can be
identified, due to the possibility to automate identification of suitable performance criteria,

monitor and evaluate performance against standards, and propose trajectories for corrective

action based on predictive analysis. In addition to the AI applications we discussed in the

“Internal analysis” subsection, particularly compelling is the use of AI project management and

internal communication. As more and more business activities are managed through increasingly

integrated digital platforms, AI will become a salient tool for supporting strategists to evaluate

strategy implementation and performance.

The Nexus of Strategizing and AI: Issues for Theory and Future Research

Future research imperatives

As mentioned above, AI may offer many attractive functions for strategists in conducting and

orchestrating strategic processes. Strategy scholars need to continue explore how such

technologies may aid in strategy analysis, formulation, and implementation. At the same time,

there is a particular need for scholars to take a prudent approach to AI and attempt to identify

boundaries as well as challenges in implementing such systems in organizations. At the time of

this writing, there is much enthusiasm expressed by scientists, consultants, technology

companies, and managers about the potential of AI, but in our view much of AI’s potential and

advantages are overstated and lack solid grounding in an understanding of the technology itself.

In the next subsection we address selected challenges for AI practice and research, but we first

offer several imperatives that we believe can guide research in the nexus between strategizing

and AI and we identify elements of a research agenda that may help strategy scholars launch

theoretical work and empirical inquiries.

First, it is key for scholars and strategists alike to understand that there is no magic bullet,

single solution, or best practice for AI applications in strategy processes. The introduction of AI
to organizational processes needs to be an iterative process that is based on trial-and-error

learning. Like the functioning of ML algorithms, managers should be open-minded when

exploring various technologies, learn from their own and others’ experiences, and continuously

improve their processes.

How strategists approach AI and how AI technology adoption unfolds in organizations

offer many research opportunities. For example, in line with earlier research in technology

adoption (King & He, 2006), scholars may ask which factors affect successful or failed

implementation of AI in strategy processes. Critical in this respect are contextual factors (e.g.,

the need for accuracy and data) that relate to strategists’ acceptance of AI, their perceptions of AI

usefulness, and their behavioral intentions to utilize the technology.

Another important factor in the adoption of AI relates to strategists’ cognitive and

affective trust in such systems. Such trust may be shaped by AI’s reliability, transparency,

immediacy, and other factors (Glikson & Woolley, 2020). As trust may be related to a host of

individual factors, as well as technological factors, it will be useful to explore what algorithms

are best suited for which specific tasks or problems in the strategy process.

As we have shown in this chapter, AI is currently used to support strategizing, but in a

somewhat piecemeal fashion. AI may be used during competitor analysis or legal analysis, or it

may be aimed at supporting acquisition target assessment or strategy controls. A question of

interest to strategy scholars is whether there exist technological architectures that can integrate

and enable specific AI applications throughout the various tasks that make up strategy processes.

A key rationale for examining this question is how to make AI more effective in supporting the

firm’s overall strategizing. This would require a study of architectural solutions, for example, by

the methods of design science, similar to studies of enterprise architectures for resource planning
(see Haki et al., 2020). For example, while an isolated prediction algorithm tuned to market

demand for certain product categories may give valuable information to the strategists, an open

question is what added insights adjacent algorithms linked to this prediction algorithm could

offer. Algorithms predicting factor costs/raw materials pricing, when linked with those

predicting demand, may provide essential competitive insights to the strategists in planning

product positioning on platforms. Scholars may want to investigate how such architectures may

generate competitively valuable information to the firm, and more generally, if and how an AI

architecture could become a new potential source of competitive advantage.

Generally, how AI can become a source of competitive advantage remains a critically

important but ambiguous question. We expect that the answer to this question will depend on

two factors: where and how AI is put to use, as well as the nature of data and the evolution of

algorithms. Specifying the first factor would demand a step-by-step examination of what types of

AI are utilized in strategy processes, and where and how they are used, and with what intensity.

In conjunction with these, researchers may aim to answer the following query: “To what extent

does AI augment and replace strategists?” The overall motivation for analyzing AI in strategy

processes is that such processes, if conducted efficiently and effectively, and they are based on

the valuable and unique identification of a firm’s strengths and weaknesses as well as novel

opportunities and threats in the environment, can be a critical source of competitive advantage

(Barney, 1991).

Specifying the second factor (the nature of data and the evolution of algorithms) would

demand a much more detailed level of scrutiny of the chosen algorithms, available training data,

and accuracy in predictions. The aim of this work would be to establish what are common and

popular algorithms in use, and what types of data become input to machine learning. The
combination of unique and rapidly updated data with efficient algorithms deployed in strategy

processes may offer the firm substantial rents compared with those that use inferior algorithms

and more constrained, dated, or flawed data; and that fail to continuously utilize AI in the

strategy process (e.g., to investigate fundamental strategic assumptions). However, unless

scholars successfully focus on this second factor connecting AI to competitive advantage, AI-

oriented strategy research may suffer the same fate as much of the study of firm performance and

information technology (e.g., Powell & Dent-Micallef, 1997), which consistently demonstrated

weak correlations. It was not until the literature shifted its focus from IT investment to IT

strategy or capabilities to build and utilize IT, that relative performance effects of IT

development and deployment could be discerned (Mithas & Rust, 2016).

As we discussed previously, modern AI drawing on big data, deep learning, or ML tools

demands high levels of investments for a continued period of time (Shrestha, Krishna, & von

Krogh, 2021). Thus, it could be beneficial for adopting firms to forge partnerships on AI

development and deployment. A large number of companies including Google, Intel, PWC,

McKinsey, IBM, Samsung, Sony, and Amazon have joined forces to form “partnership in AI”

with the sole purpose of shaping best practices, research, and public dialogue about AI’s benefits

for business and society. Strategy scholars could investigate the drivers, processes, and outcomes

of these partnerships and how they shape the adoption of AI in strategy processes.

Moreover, companies thinking of integrating AI applications in their strategy process are

required to build an AI team composed of programmers and data scientists. Their knowledge on

algorithms and its functioning is to be complemented with domain expertise, which requires the

design of team structures that best benefit such complementarities. Scholars of strategy,
organization theory, and organizational behavior can attempt to examine such team structures,

antecedents for AI team performance, and team processes.

Finally, the rapid emergence of AI in the strategy process may have significant

consequences for human creativity and ingenuity inextricably linked to successful process

outcomes, and the politics surrounding corporate strategy. It is as of yet difficult to understand

such consequences, which call for research from a broad set of perspectives rooted in

organization behavior, organization theory, information systems, and strategic management. For

example, with the increasing use of AI in strategic analysis, strategists may feel less inclined to

introduce creative elements, such as novel and non-traditional data sources or radically different

interpretations in strategic analysis. Moreover, those who control the deployment of algorithms

and interpret analytical outcomes may also gain unprecedented power in the novel approach to

strategizing. It will be critically important for scholars to investigate such power dynamics.

Finally, future research should explore the risks associated with reliance on AI in new venture

activities, including the implications for the venture’s ability to learn and develop critical

capabilities. This is key given the limited transparency that is currently inherent to many ML

systems (Castelvecchi, 2016).

Besides these research questions, ML algorithms also generates useful avenues to

investigate how algorithm-supported methods can be developed to conduct strategy research (He

et al., 2020, Shrestha et al., forthcoming).

Challenges and risks for strategizing

The application of AI for strategy analysis could lead to various social and organizational

challenges. First, the application of AI requires the collection of vast amounts of personal and

interactional data (e.g., on political sentiments or consumer behavior), which often interferes
with individual privacy and data security. Firms should be highly cautious and aware of potential

risks when using such data for strategic analysis. In particular, when using customer, employee,

or client data for strategic analysis, firms must ensure that regulations or their code of conduct on

privacy are not being violated unknowingly.

In recent years, the media have drawn the public’s attention to various such high-profile

privacy violations. Triggered by such cases, policy makers, investors, and the public are

sensitized to the (mis)use of private information. As a consequence, various privacy regulations

such as the General Data Protection Regulation (GDPR)––the European Union’s regulation on

data protection and privacy––will continue to be put in place, accompanying new standards on

an individual’s right to their own information. In addition to fulfilling ethical responsibilities to

stakeholders regarding their information, firms will want to avoid hefty fines and/or reputational

harm stemming from violations.

Second, with modern ML algorithms, information in the form of digital bits can be

searched and replicated from terabytes of data at almost zero cost. With the reduced search,

tracking, and verification costs in information acquisition, organizations of diverse size and

scope have relatively similar access to information, reducing information asymmetry (Goldfarb

& Tucker, 2019). As information searching forms an important constituent of strategy analysis,

such lowered costs could have potentially mixed effects. On the one hand, they make it easier to

rapidly find rare and potentially valuable strategic decision alternatives, resulting in the addition

of variety in the alternative set (Yang, 2013; Zhang et al., 2018). On the other, search algorithms

(such as recommender systems) represent popularity bias, resulting in lack of variety with

respect to information acquisition (Fleder & Hosanagar, 2009).

 Third, and critically important, ML algorithms have been shown to produce results that

are systematically prejudiced with respect to gender and ethnicity, among other factors. There

are valid concerns that use of algorithms amplifies and perpetuates social inequalities (Barocas,

Selbst, 2016; Starr, 2014). Scholars have identified various types of biases produced by

algorithms, and the inherent trade-off between algorithm performance and algorithmic fairness

makes correcting bias issues quite difficult (Shrestha, Yang, 2019; Hardt, Price, Srebro, 2016;

Kleinberg, Mullainathan, Raghavan, 2017). For instance, ML algorithms are trained on historic

data, which can reflect social inequalities or biases contained in past human decisions. Flawed

data collection or sampling procedures (either intentional or unintentional from the designer’s

perspective) could also lead to biases. Research in ML has also associated biases with the

selection of the objective function or the algorithm itself (Arduini et al., 2020; Mehrabi et al.,

2019). Thus, strategists are required to stay up-to-date with research on fair AI. Scholars suggest

that the use of more transparent and interpretable AI models could assist in exposing and

mitigating bias in algorithmic outcomes (Dodge et al., 2020). Combining domain expertise with

AI by “human-in-the loop” decision making has also been shown to reduce algorithmic bias

(Holstein et al., 2019). In order to avoid detrimental consequences, strategists are well advised to

put into place governance structures to deal with any unintentional biases produced by AI

algorithms.

Fourth, as ML algorithms learn from past data to predict the future, their efficacy is

contingent on the stability of data generation. That means, ML aimed at population prediction

works well with relatively stable phenomena, such as consumer preferences, borrowing patterns,

and demographic characteristics. Thus, relying on algorithmic predictions under conditions of

disruption, crisis, or turbulence is misguided. Under these conditions the learning accumulated
by algorithms loses relevance. Yet, when integrated into strategic processes, such as automated

analysis of market demand, it may be difficult to “decouple” the algorithm from other strategy

work. Moreover, putting in place an algorithm that performs complex learning with the intent to

inform strategists, demands continuous data replenishment. Data flows may easily be interrupted,

as we have seen with the disruptions in telecom infrastructures and other grids (e.g., energy). In

general, it is worthwhile to recall that an algorithm’s ability to make powerful predictions is only

as strong as the quality of the data fed to it and the extent to which it fits the processing task. The

strategist will likely be much worse off if she draws on flawed predictions based on low-quality

data, than if she uses her own good judgment.

Fifth, algorithms often feed off each other’s behavior. Output from one algorithm often

impacts the learning of another. This interconnectedness may sometimes lead to erratic behavior

as we have seen in the case of high-frequency trading, where algorithms have put significant

financial assets at risk. Thus, in a market or industry where competitors are increasingly utilizing

algorithms, their deployment demands careful human scrutiny, governance, and the enforcement

of behavioral rules (e.g., a limit to volume of assets exposed in high-frequency trading).

Consider, for instance, the case of algorithms used for pricing, which are often mis-specified and

ignore effects that are outside their own control (e.g., market and competitors’ prices). Research

has shown that when competing sellers use ML for real-time dynamic pricing, it leads to

overestimation of costs and market collusion, resulting in inflated prices (Harrington, 2018).

Finally, the use of AI-based technologies for strategic analysis requires and facilitates

maintaining information within organizations in a central information system, cloud, or digital

platform that can then be easily processed and used to train the predictive models. Such collected

digital information when strategically relevant, demands protective measures (Liebeskind, 1996),
which in the case of modern digital technologies can become very costly. Traditionally,

strategists have often held privileged access to information and expertise. With these

technologies, maintaining such privileged access becomes increasingly costly.

Conclusion

We showed that AI holds great promise for the work of strategists and the effective and efficient

conduct of strategic processes. The next few years will likely see a rapid adoption of ML

algorithms that fundamentally alter the way we strategize. This opens up many important

research opportunities for strategy scholars, some of which we discussed here. As practice begins

to struggle with smart technologies, managers will urgently need our research findings to guide

them.

References
Agrawal S, Chatterjee K, Novotný P. 2018. Lexicographic ranking supermartingales: an efficient

approach to termination of probabilistic programs. Proceedings of the ACM on Programming

Languages 2(POPL): 1–32. [Link]

Arduini M, Noci L, Pirovano F, Zhang C, Shrestha YR, Paudel B. 2020. Adversarial Learning

for Debiasing Knowledge Graph Embeddings. In MLG 2020: 16th International Workshop on

Mining and Learning with Graphs - A Workshop at the KDD Conference, August 24, 2020,

San Diego, [Link], New York, NY, USA: 7.

Avramov D, Cheng S, Metzker L. 2019. Machine Learning versus Economic Restrictions:

Evidence from Stock Return Predictability. SSRN Electronic Journal.

[Link]

Bail CA. 2014. The cultural environment: measuring culture with big data. Theory and Society

43(3/4): 465–482. [Link]

Barney J. 1991. Firm resources and sustained competitive advantage. Journal of Management

17(1): 99–120. [Link]

Barocas S, Selbst AD. 2016. Big data’s disparate impact. SSRN Electronic Journal.

[Link]

Baum JR, Wally S. 2003. Strategic decision speed and firm performance. Strategic Management

Journal 24(11): 1107–1129. [Link]

Biessmann F, Lehmann P, Kirsch D, Schelter S. 2016. Predicting political party affiliation from

text. PolText, 14–19. [Link]

Brynjolfsson E, Mitchell T. 2017. What can machine learning do? Workforce implications.

Science 358(6370): 1530–1534. [Link]

Castelvecchi D. 2016. Can we open the black box of AI? Nature 538(7623): 20–23.

[Link]

Chalmers D, MacKenzie NG, Carter S. 2020. Artificial intelligence and entrepreneurship:

implications for venture creation in the Fourth Industrial Revolution. Entrepreneurship

Theory and Practice. [Link]

Chan JT, Zhong W. 2018. Reading China: Predicting Policy Change with Machine Learning.

SSRN Electronic Journal. [Link]

Cheung CF, Wang WM, Lo V, Lee WB. 2004. An agent-oriented and knowledge-based system

for strategic e-procurement. Expert Systems 21(1): 11–21. [Link]

0394.2004.00259.x

Cielen A, Peeters L, Vanhoof K. 2004. Bankruptcy prediction using a data envelopment analysis.

European Journal of Operational Research 154(2): 526–532.

[Link]
Cockburn I, Henderson R, Stern S. 2018. The impact of artificial intelligence on innovation.

NBER Working Paper Series 24449. National Bureau of Economic Research, Cambridge,

MA. [Link]

Coletto M, Lucchese C, Orlando S, Perego R. 2015. Electoral Predictions with Twitter: A

Machine-Learning Approach. CEUR Workshop Proceedings. Retrieved from

[Link]

Davenport TH, Ronanki R. 2018. Artificial intelligence for the real world. Harvard Business

Review 96(1): 108–116. [Link]

Dodge J, Gururangan S, Card D, Schwartz R, Smith NA. 2020. Show your work: improved

reporting of experimental results. EMNLP-IJCNLP 2019: 2019 Conference on Empirical

Methods in Natural Language Processing and 9th International Joint Conference on Natural

Language Processing, Proceedings of the Conference, 2185–2194.

[Link]

Doyle G, Srivastava SB, Goldberg A, Frank MC. 2017. Alignment at work: using language to

distinguish the internalization and self-regulation components of cultural fit in organizations.

ACL 2017--55th Annual Meeting of the Association for Computational Linguistics,

Proceedings of the Conference (Long Papers) 1: 603–612. [Link]

1056

Ferràs-Hernández X, Tarrats-Pons E, Arimany-Serrat N. 2017. Disruption in the automotive

industry: a Cambrian moment. Business Horizons 60(6): 855–863.

[Link]

Fethi MD, Pasiouras F. 2010. Assessing bank efficiency and performance with operational

research and artificial intelligence techniques: a survey. European Journal of Operational

 Research 204(2): 189–198.

[Link]

Fleder D, Hosanagar K. 2009. Blockbuster culture’s next rise or fall: the impact of recommender

systems on sales diversity. Management Science 55(5): 697–712.

[Link]

Franz L, Shrestha YR, Paudel B. 2020. A Deep Learning Pipeline for Patient Diagnosis

Prediction Using Electronic Health Records. In BioKDD 2020: 19th International Workshop

on Data Mining in Bioinformatics, August 24, 2020, San Diego [Link], New York, NY,

USA

Glikson E, Woolley AW. 2020. Human trust in artificial intelligence: review of empirical

research. Academy of Management Annals 14(2), 627–660.

[Link]

Goldfarb A, Tucker C. 2019. Digital economics. Journal of Economic Literature 57(1): 3---43.

Gomez-Uribe CA, Hunt N. 2015. The Netflix recommender system: algorithms, business value,

and innovation. ACM Transactions on Management Information Systems 6(4): 1–19.

[Link]

Haefliger S, Jäger P, von Krogh G. 2010. Under the radar: industry entry by user entrepreneurs.

Research Policy 39(9): 1198–1213. [Link]

[Link]

Haenlein M, Kaplan A. 2019. A brief history of artificial intelligence: on the past, present, and

future of artificial intelligence. California Management Review 61(4): 5–14.

[Link]

Haki K, Beese J, Aier S, Winter R. 2020. The evolution of information systems architecture: an
 agent-based simulation model. Management Information Systems Quarterly 44(1).

[Link]

Hamel G, Prahalad CK. 1994. Competing for the Future. Harvard Business School Press.

[Link]

Hamet P, Tremblay J. 2017. Artificial intelligence in medicine. Metabolism: Clinical and

Experimental 69: S36–S40. [Link]

Hardt M, Price E, Srebro N. 2016. Equality of opportunity in supervised learning. Advances in

Neural Information Processing Systems, 3323–3331. [Link]

Harrington JE. 2018. Developing competition law for collusion by autonomous artificial agents.

Journal of Competition Law and Economics 14(3), 331–363.

[Link]

He VF, Puranam P, Shrestha YR, von Krogh G. 2020 Resolving governance disputes in

communities: A study of software license decisions Strategic Management Journal, 41 (10):

1837-1868

Heinonen K. 2011. Consumer activity in social media: anagerial approaches to consumers’ social

media behavior. Journal of Consumer Behaviour 10(6): 356–364.

[Link]

Hitt MA, Ireland RD, Hoskisson RE. 2015. Strategic Management: Competitiveness and

Globalization--Concepts and Cases. Cengage Learning Asia Pte Ltd.

Hollebeek LD, Conduit J, Brodie RJ. 2016. Strategic drivers, anticipated and unanticipated

outcomes of customer engagement. Journal of Marketing Management 32(5–6): 393–398).

[Link]

Holstein K, Wortman Vaughan J, Daumé H, Dudik M, Wallach H. 2019. Improving fairness in

 machine learning systems: what do industry practitioners need? Proceedings of the 2019 CHI

Conference on Human Factors in Computing Systems, 1–16.

[Link]

Homburg C, Bucerius M. 2006. Is speed of integration really a success factor of mergers and

acquisitions? An analysis of the role of internal and external relatedness. Strategic

Management Journal 27(4): 347–367. [Link]

Huang AC, Jiang T, Liu YX, Bai YC, Reed J, Qu B, Goossens A, Nützmann HW, Bai Y,

Osbourn A. 2019. A specialized metabolic network selectively modulates Arabidopsis root

microbiota. Science 364(6440). [Link]

Humphreys P, McIvor R, Huang G. 2002. An expert system for evaluating the make or buy

decision. Computers & Industrial Engineering 42(2): 567–585.

[Link]

Ireland RD, Hitt MA, Sirmon DG. 2003. A model of strategic entrepreneurship: the construct

and its dimensions. Journal of Management 29(6): 963–989. [Link]

2063_03_00086-2

Jarrahi MH. 2018. Artificial intelligence and the future of work: human---AI symbiosis in

organizational decision making. Business Horizons 61(4): 577–586.

[Link]

Jarzabkowski P, Balogun J, Seidl D. 2007. Strategizing: The challenges of a practice perspective.

Human Relations 60(1): 5–27. [Link]

Jean N, Burke M, Xie M, Davis WM, Lobell DB, Ermon S. 2016. Combining satellite imagery

and machine learning to predict poverty. Science 353(6301): 790–794.

[Link]
Keding C. 2020. Understanding the interplay of artificial intelligence and strategic management:

four decades of research in review. Management Review Quarterly, 1–44.

[Link]

King WR, He J. 2006. A meta-analysis of the technology acceptance model. Information and

Management 43(6): 740–755. [Link]

Kleinberg J, Mullainathan S, Raghavan M. 2017. Inherent trade-offs in the fair determination of

risk scores. Leibniz International Proceedings in Informatics (LIPIcs), 67.

[Link]

Kou G, Chao X, Peng Y, Alsaadi FE, Herrera-Viedma E. 2019. Machine learning methods for

systemic risk analysis in financial sectors. Technological and Economic Development of

Economy 25(5): 716–742. Vilnius Gediminas Technical University.

[Link]

Kshirsagar V, Wieczorek J, Ramanathan S, Wells R. 2017. Household poverty classification in

data-scarce environments: a machine learning approach. ArXiv: Machine Learning.

[Link]

Kucukkeles B, Ben-Menahem SM, von Krogh G. 2019. Small numbers, big concerns: practices

and organizational arrangements in rare disease drug repurposing. Academy of Management

Discoveries 5(4): 415–437. [Link]

Lawrence T. 1991. Impacts of artificial intelligence on organizational decision making. Journal

of Behavioral Decision Making 4(3): 195–214. [Link]

Lee C, Kwon O, Kim M, Kwon D. 2018. Early identification of emerging technologies: a

machine learning approach using multiple patent indicators. Technological Forecasting and

Social Change 127: 291–303. [Link]

Liebeskind JP. 1996. Knowledge, strategy, and the theory of the firm. Strategic Management

Journal 17(S2), 93---107.

March JS, Simon HA. 1958. Organizations. Wiley: New York.

McCarthy J. 1981. Epistemological problems of artificial intelligence. In Readings in Artificial

Intelligence, Webber BL, Nilsson NJ (eds). Elsevier: 459–465. [Link]

0-934613-03-3.50035-0

McCorduck P. 2004. Machines Sho Think: A Personal Inquiry into the History and Prospects of

Artificial Intelligence. AK Peters/CRC Press.

Mehrabi N, Morstatter F, Saxena N, Lerman K, Galstyan A. 2019. A survey on bias and fairness

in machine learning. ArXiv, abs/1908.0.

Menon A, Choi J, Tabakovic H. 2018. What you say your strategy is and why it matters: natural

language processing of unstructured text. Academy of Management Proceedings, 2018(1).

[Link]

Mithas S, Rust R. 2016. How information technology strategy and investments influence firm

performance: conjecture and empirical evidence. Management Information Systems Quarterly

40(1). [Link]

Nag R, Hambrick DC, Chen MJ. 2007. What is strategic management, really? Inductive

derivation of a consensus definition of the field. Strategic Management Journal 28(9): 935–

955. [Link]

Nissen ME, Sengupta K. 2006. Incorporating software agents into supply chains: experimental

investigation with a procurement task. MIS Quarterly: Management Information Systems

30(1): 145–166. [Link]

Orlikowski W J, Scott SV. 2008. Sociomateriality: challenging the separation of technology,

 work and organization. The Academy of Management Annals 2(1): 433–474.

[Link]

Pant G, Sheng ORL. 2015. Web footprints of firms: using online isomorphism for competitor

identification. Information Systems Research 26(1): 188–209.

[Link]

Powell TC, Dent-Micallef A. 1997. Information technology as competitive advantage: the role of

human, business, and technology resources. Strategic Management Journal 18(5): 375–405.

[Link]

Ranaei S, Karvonen M, Suominen, A., Kässi, T. (2016). Patent-based technology forecasting:

case of electric and hydrogen vehicle. International Journal of Energy Technology and Policy

12(1): 20–40. [Link]

Randall C, Dent EB. 2019. Reconciling the historical divide between strategy process and

strategy content. Journal of Management History 25(3): 401–427.

[Link]

Reeves M, Levin S, Ueda D. 2016. “The Biology of Corporate Survival.” BCG Henderson

Institute. [Link]

Robinson CV, Ahmad F, Simmons JEL. 2020. Consolidation and fragmentation in

environmental scanning: a review and research agenda. Long Range Planning.

[Link]

Shrestha YR, Ben-Menahem SM, von Krogh G. 2019. Organizational decision-making structures

in the age of artificial intelligence. California Management Review 61(4): 66–83.

[Link]

Shrestha YR, He VF, Puranam P, von Krogh G. Forthcoming. Algorithmic induction through
 machine learning: opportunities for management and organization research. Organization

Science

Shrestha YR, Krishna, V von Krogh, G. 2021 Augmenting Organizational Decision-Making with

Deep Learning Algorithms: Principles, Promises, and Challenges. Journal of Business

Research, 123, 588-603

Shrestha YR, Yang Y. 2019. Fairness in algorithmic decision-making: Applications in multi-

winner voting, machine learning, and recommender systems Algorithms, 12 (9): 199

Simon HA. 1947. Administrative Behavior: A Study of Decision-making Processes in

Administrative Organization. Macmillan: New York.

Starr P. 2014. Is the past in our future? Contemporary Sociology: A Journal of Reviews 43(6),

795–800. [Link]

Strohmeier S, Piazza F. 2013. Domain driven data mining in human resource management: a

review of current research. Expert Systems with Applications 40(7): 2410–2420.

[Link]

Suominen A, Toivanen H, Seppänen M. 2017. Firms’ knowledge profiles: mapping patent data

with unsupervised learning. Technological Forecasting and Social Change 115: 131–142.

[Link]

Treleaven P, Batrinca B. 2017. Algorithmic regulation: automating financial compliance

monitoring and regulation using AI and blockchain. Journal of Financial Transformation 45:

14–21. [Link]

von Krogh G. 2018. Artificial intelligence in organizations: new opportunities for phenomenon-

based theorizing. Academy of Management Discoveries 4(4): 404–409.

[Link]
von Krogh G, Roos J. 1995. Organizational Epistemology. St. Martin’s Press: New York;

Macmillan: Basingstoke.

Whittington, R. (2014). Information systems strategy and strategy-as-practice: a joint agenda.

Journal of Strategic Information Systems 23(1): 87–91.

[Link]

Wu C-F, Huang SC, Chang T, Chiou C-C, Hsueh H-P. 2020. The nexus of financial

development and economic growth across major Asian economies: evidence from bootstrap

ARDL testing and machine learning approach. Journal of Computational and Applied

Mathematics 372: 112660. [Link]

Yang H. 2013. Targeted search and the long tail effect. The RAND Journal of Economics 44(4):

733–756. [Link]

Yoo PD, Kim MH, Jan T. (2005). Machine learning techniques and use of event information for

stock market prediction: a survey and evaluation. In International Conference on

Computational Intelligence for Modelling, Control and Automation and International

Conference on Intelligent Agents, Web Technologies and Internet Commerce (CIMCA-

IAWTIC’06) 2: 835–841. [Link]

Yoo S, Digman LA. 1987. Decision support system: a new tool for strategic management. Long

Range Planning 20(2): 114---124.

Yousfi-Monod M, Farzindar A, Lapalme G. 2010. Supervised machine learning for summarizing

legal documents. Canadian AI 2010. In Lecture Notes in Artificial Intelligence, vol. 6085,

Farzindar A, Keselj V (eds). Springer-Verlag: Berlin, Heidelberg: 51–62.

[Link]

Zafarani R, Abbasi MA, Liu H. 2014. Social Media Mining: An Introduction. Cambridge
 University Press. [Link]

Zhang S, Ke X, Frank Wang XH, Liu J. 2018. Empowering leadership and employee creativity:

a dual-mechanism perspective. Journal of Occupational and Organizational Psychology

91(4): 896–917. [Link]

Table 1: Strategists’ core activities and AI augmentation in the strategy process

Stage of strategy process Strategic analysis Strategy formulation and

implementation
Strategists’ core activities External analysis of Formulating corporate and business
macro environment strategy
(e.g., PESTEL, which
refers to Political
Economic Socio-cultural
Technological
Environmental and
Legal)
External analysis of Creating and selecting strategic
competitive dynamics options (e.g., mergers and
acquisitions [M&As], cooperative
strategies)
Internal analysis of Strategic innovation,
human, financial, and entrepreneurship, and renewal
auxiliary resources and
capabilities
Measuring and evaluating
implementation and performance of
strategic actions
AI augmentation Continuous interaction Model business portfolio
with external performance, synergies, and
environment to forage portfolio risks
information from
external data sources
Interpreting data Screen for novel and unobvious
through identifying strategic options (e.g., M&As,
structures and patterns collaboration partners, innovation
in high-dimensional data projects)
Predictive analysis Increase speed of decision making
Monitor strategy implementation
on online platforms and digital
communication channels
Challenges and risks Data privacy and data Limited variety and creativity in
security strategic options
Data biases and lack of Perpetuating social, political,
data variety economic equalities
Interruptions to data Erratic strategic actions resulting
flows from erratic data and analysis
Cost of protecting
access to privileged
information and
expertise