Content uploaded by Arindra Nath Mishra
Author content
All content in this area was uploaded by Arindra Nath Mishra on Jan 11, 2023
Content may be subject to copyright.
Business value appropriation
roadmap for artificial intelligence
Arindra Nath Mishra and Ashis Kumar Pani
Department of Business Management, XLRI, Jamshedpur, India
Abstract
Purpose –Artificial intelligence (AI) is deemed to have a significant impact as a value driver for the firms
and help them get an operational and competitive advantage. However, there exists a lack of understanding of
how to appropriate value from this nascent technology. This paper aims to discuss the approaches toward
knowledge and innovation strategies to fill this gap.
Design/methodology/approach –The discussion presents a review of the extant strategy and
information systems literature to develop a strategy for organizational learning and value appropriation
strategy for AI. A roadmap is drawn from ambidexterity and organizational learningtheories.
Findings –This study builds the link between learning andambidexterity to propose paths for exploration
and exploitation of AI. The study presents an ambidextrous approach toward innovation concerning AI and
highlights the importance of developing as well as reusing the resources.
Research limitations/implications –This study integrates over three decades of strategy and
information systems literature to answer questions about value creation from AI. The study extends the
ambidexterity literature with contemporary.
Practical implications –This study could help practitioners in making sense of AI and making use of
AI. The roadmap could be used as a guide for the strategy development process.
Originality/value –This study analyzes a time-tested theoretical framework and integrates it with
futuristic technology in a way that could reduce the gap between intent and action. It aims to simplify the
organizational learning and competency development for an uncertain, confusing and new technology.
Keywords Ambidexterity, Business transformation, Organizational learning,
New product development, Artificial intelligence, E-business strategy
Paper type General review
1. Introduction
Artificial intelligence (AI) is an attempt to understand and build a machine capable of doing
intelligent tasks. AI could be pivotal to the second machine age (Brynjolfsson and McAfee,
2016) and help us in mastering our physical and intellectual environment, leading to
prosperity for the humankind. We have dabbled in creating a “thinking machine”for several
decades. However, recent headway into AI encourages us to seriously consider the projected
impact on the world (Kurzweil et al., 1990;Kurzweil, 2006;Agrawal et al.,2017). This
burgeoning transformation of our world hints at an immense opportunity for firms. It has
been discussed that companies could use AI for optimizing the business departments like
operations (Baryannis et al., 2019), marketing (Roos and Kern, 1996;Sterne, 2017) and
human resources (HR) (Sivathanu and Pillai, 2018;Stone et al.,2018). AI adoption promises
value embedded in its untapped potentials and warrants a detailed investigation (Mckinsey.
com, 2018).
Despite the widespread understanding of the potential of AI, it is not fully understood,
which prevent firms from implementing AI-based technology solutions andextract business
value from them. Ransbotham et al. (2017) reported that about 85 per cent of the executives
believe AI can help them in business. However, only5 per cent of companies had extensively
Business value
353
Received 12 July2019
Revised 30 January2020
Accepted 1 March2020
VINE Journal of Information and
Knowledge Management Systems
Vol. 51 No. 3, 2021
pp. 353-368
© Emerald Publishing Limited
2059-5891
DOI 10.1108/VJIKMS-07-2019-0107
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/2059-5891.htm
incorporated AI, and 20 per cent had partially used AI. This points toward a gap between
potential and actual adoption. Another survey found that while 83 per cent banks had
considered using AI or machine learning in their business, the implementation rate was
around 67 per cent (Dougal, 2018). The gap has been explained by Dougal as a lack of
knowledge about the application of these technologies into business problems. The
knowledge would come through organizational learning, post which value can be extracted.
We have differentiated process into three steps –learning, research and development (R&D)
and value creation. This exploration is guided by three pertinent research questions (RQs)
around deriving business value from AI:
RQ1. How could organizations develop knowledge in AI?
The approach toward developing technological competencies required to appropriate value
from AI would come from organizational learning (OL) (Crossan and Berdrow, 2003;Real et
al.,2006). These competencies could be developed through the exploration of new avenues
for AI developmentas well as making exploitative use of AI by incremental innovation:
RQ2. What are the ways of deriving value from AI for an organization?
An organization can exploit extant knowledge as well as explore new ways of using AI as
proposed by March (1991). However, extending the extant discussion on value creation in
current technological context could potentially help formulate a robust strategy for delivery
value:
RQ3. What are some themes of current uses of AI in businesses?
Some of the AI-based tools are part of information systems in organizations and could help
in improving the processes. If we can highlight how it has impacted departments like
marketing and operations, we could show the potential use cases of AI-based technologies in
process improvement.
2. Artificial intelligence
We live in a post-industrial, information-based society (Duff, 2004;Webster, 2007). Over the
past four decades, information systems have moved from a peripheral to a central one.
Transitioning from decision support systems to data warehouse to real-time warehouse to
big data analytics, we are eventually moving toward the “cognitive-computing”era
(Watson, 2017). AI will drive this new wave of information technology (IT) at the core of
information-based systems, thus opened up avenues for the use of AI in analytics, language
processing and visual processing. AI could benefit humanity through a massive increase in
information processing accuracy and efficiency, unlocking economic and social
development (Hall and Pesenti, 2017). Some of the recent applications have been to chatbots
(Hill et al., 2015), translation services (Bahdanau et al.,2014;LeCun et al.,2015;Wu et al.,
2016;Johnson et al.,2017), robotics and autonomous agents (Dirican, 2015;Reitman, 1984;
Tirgul and Naik, 2016) and virtual assistants like OK Google, Siri, Alexa and Cortana
(Bushnel, 2018). It is interpolating from these developments that it can be estimated that in
the future, machines could replace humans in tasks like driving cars, solving problems or
managing logistics (Brynjolfsson and McAfee, 2016). This is achieved by mimicking the
way we humans learn. Tasks like reading a newspaper, is essentially recognizing the letters,
assembling them into meaningful words and sentences. AI can do the same using
algorithms that are broadly called machine learning (ML) algorithms. ML uses statistics for
finding patterns in huge amounts of data (Hao, 2018).
VJIKMS
51,3
354
AI makes it possible for machines to be able to adapt and solve problems in uncertain
domains. However, the way AI may acquire knowledge and makes sense of the world could
shape its perception of the world (Sanzogni et al.,2017). This would require careful
deliberation of sources and boundaries of knowledge within AI.
3. Classical technology dilemma: exploration vs exploitation
According to March (1991), there are two broad ways to generate value from technology:
firstly, through exploitation, which is “extension of existing competencies, technologies, and
paradigms,”while exploration is about finding new alternatives. Factors like uncertainty in
the results, time of development, novelty of the development compared to the current
process could help decide which path to take (Kuittinen et al., 2013). However, over-reliance
on one of these strategies could be detrimental as it leads to “competency trap”; hence, it is
suggested to strike a balance between both (Liu, 2006). Ambidextrous organizations are
those that attempt to balance both (Tushman and O’Reilly, 1996;O’Reilly and Tushman,
2004).
This means that organizations need to have skill sets to survive in mature markets,
which play by the rules of efficiency, and new products markets, which play by the rules of
innovation, agility and flexibility (Tushman and O’Reilly, 1996). Though there exist
competing schools of thoughts who prescribe focusing on one or the other or both, we look
into value creation as a function of ambidexterity as it has been shown to work best for
survival of new ventures (Hill and Birkinshaw, 2014).
While the competencies and the R&D are integral components of any OL strategy, the
major rethinking for AI would be based on the two aspects that have come up in recently:
servicitization and open development. Servicitization is turning products into service where
possession of physical hardware is with the purveyor who lets the buyers subscribe to them
without the hassle and cost of procurement (Ramiller et al.,2008). One of the ways to achieve
servicitization is technology development over a platform.
On the other hand, open development, as the name suggests, involves a non-proprietary
approach of managing intellectual property. Open innovation has also been termed as
external knowledge exploration (Lichtenthaler, 2011). This can come in the form of
outsourcing, sharing and collaboration, thereby developing competencies. Open innovation
is an emergent method of R&D (Enkel et al.,2009). We have tried to explore the relevance of
open innovation in AI R&D.
4. Roadmap for artificial intelligence
We build our roadmap from the empirical findings of Turulja and Bajgori
c (2018). They
demonstrated that “knowledge”is the precursor to “innovation,”in turn leads to business
performance in firms. This broad knowledge innovation strategy can be played out by
several combinations, as shown in Figure 1.
There are four different pathways presented, two each for type of knowledge strategy
and two types of innovation strategies. As discussed earlier, two approaches to knowledge
could be either to apply existing knowledge (exploitation) to solve business problems or seek
new knowledge (exploration) to address business problems (March, 1991). The goal of all
business processes is to convert inputs into valuable outputs. New technology is expected to
improve these processes through innovation. Business value can be created either through
new products (product innovation) or through optimizing the business process (process
innovation). In the context of AI, the innovation in product comes through new product
development (NPD). Another type of innovation would be innovation in process which
would be brought about by business process transformation (BPT).
Business value
355
The technology transformation roadmap is indicated as an approach toward business value
appropriation from AI as it moves from its current to future state. The current state is the
one where the firm has not yet embraced AI for its business, while the future state is the one
where it would have embraced and used AI. The three steps of OL, R&D and value creation
are not strictly sequential neither chronological. However, Figure 2 depicts a general
direction toward the future state, which is dependent on the organizational learning, which
would be elemental for the R&D process that would be used for the value creation. This is
seen through the dimensions of either explorationor exploitation of technology.
4.1 Organizational learning
OL is defined as “development of insights, knowledge, and associations between past
actions, the effectiveness of those actions, and future actions”(Fiol and Lyles, 1985). OL is a
continuous, irreversible and path-dependent process (Nieto, 2004), which is a process of
encoding learning into a guiding framework for the organization (Levitt and March, 1988).
One of the critical aspects of OL is organizational knowledge (OK). OK is the ability to carry
out business processes based upon the past collective understanding of the domain and
context (Tsoukas and Vladimirou, 2001). It has been shown that OK can serve both as a
Figure 1.
KM, innovation and
performance
Figure 2.
Technology
transformation
roadmap
VJIKMS
51,3
356
barrier as well as a source of inspiration for the NPD (Carlile, 2002). On the other hand, there
is a strong link between success of BPT and knowledge management practices (Jang et al.,
2002).
Exploitation is essentially knowledge application and it gets operationalized in sensing
and seizing new opportunities for the firm (Teece, 1998). Exploitation strategy works like an
adaptation that can be done quickly and predictable, while exploration has inherent risk,
and it is a time-taking process (Real et al., 2006). This could be more suited for firms that
needs to implement fast and frugal AI innovation. On the other hand, the exploitation
strategy works through knowledge acquisition and represents the firm’s ability to recognize
usable knowledge, then endeavor to absorb it (Liao et al., 2009). The development of a
proprietary algorithm (one that is more accurate in prediction or faster or computationally
efficient than previous algorithms) will lead to a distinctive competency. These types of
distinctive competencies (Real et al.,2006) help the organization in developing an inimitable
value proposition for the customers, leading to a competitive advantage (Crossan and
Berdrow, 2003).
In the case of NPD, primarily, there needs to be a focus on exploration. On the other hand,
for BPT, exploitation priority is a better approach. However, in both cases, the
complementary approach should also take place to balance out the benefits.
The takeaway for a firm indulging in BPT is that IS competencies lead to increased
entrepreneurial agility for the firm (Chakravarty et al.,2013), which in turn leads to higher
firm performance (Sambamurthy et al., 2003).
4.2 Research and development
Each company may have different requirement and needs, which would result in differing
information systems requirements (Gregory, 1995). R&D serves the core function of
exploration, and we will have a look at different kinds of R&D. There are three kinds of
R&D setups: decentralized, network and integrated (DeSanctis et al.,2002). Decentralized
designs are more suitable for firms that want to improve the existing product. This would be
better for firms that are indulging in BPT using AI. It would also be optimum for a firm that
is AI producer doing NPD if AI is one of the many products it makes because of the
reusability prospects of the research insights.
4.2.1 Open innovation, outsourcing and collaboration. One of the exploitation methods
could be through outsourcing of the process to save time and reduce uncertainty (Kuittinen
et al., 2013). This is supported by an increasing trend toward external resources is upheld by
the value in sharing the development, deployment and maintenance. Open innovation is an
approach suited to such business environments with marked uncertainty in resources like
AI. AI applications can be developed using components like codes, open APIs, JSON streams
and algorithms from easily available open sources. It has been rightly said that “Most
innovations fail. And companies that don’t innovate die”(Chesbrough, 2006). The solution
proposed by Chesbrough is open innovation where a firm uses both internal and external
ideas that can result in faster product-to-market, lower cost and higher firm sustainability. It
was estimated that there were less than 10,000 professionals with the right skill set for AI
development (Nott, 2018), which leads to a skyrocketing in the salaries of AI professionals
(Metz, 2017). In this landscape, it makes even more sense to have open innovation and
indulge in collaborative development.
The second approach toward development is to outsource. Outsourcing is taking
external help in filling the gaps in an organization’s IS capabilities. These gaps are a
function of the resource attributes and resource allocation (Cheon et al., 1995). AI’s
complexity and resource dependencies call for greater need for outsourcing. It has also been
Business value
357
shown to reduce effort, increase profitability and reduce risks (Choi et al.,2018;Loh and
Venkatraman, 1995). Collaboration can also be a mechanism for the acquisition of external
resources. Universities and independent research laboratories can be of help in exploration
of knowledge for firms. It has been demonstrated by Paunov et al. (2019) that corporate
patent filings had higher citations of university publications that were located near. The
best examples are Google, Microsoft and Uber who have collaborated with universities in
Canada and the USA to co-develop AI-based technological solutions. Another successful
example is the Catapul program in UK that helps link industry and research (Davis, 2015).
On the other hand, there are examples like Australia that has low collaboration between
industry and the academia (OECD, 2019). The willingness to collaborate could be lower due
to the gap between academic and practical perspective or due to lack of direct benefits for
the practitioners (Rodríguez et al., 2014).
4.3 Value creation processes
The first decision should be prioritizing on either acquisition or application of knowledge or
a mix of both. Second would be the way knowledge is used to create value. This requires
R&D activities, which result in innovation. The knowledge processes can lead to two kinds
of innovations –product innovation resulting in new product development or process
innovation leading to an optimized process. Knowledge has been shown to leads to
innovation, which in turn leads to firm performance (Turulja and Bajgori
c, 2018).
4.3.1 New product development. One of the ways of value appropriation is NPD where
the firm may indulge in the identification of opportunities where it can enter the market with
a new offering. As discussed earlier, exploitation approach deals with the incremental
innovation, while explorative approach deals with the disruptive innovation. Incremental
innovation approach would improve a business process, while disruptive would be akin to a
new process (Norman and Verganti, 2014). However, IT projects are generally complex and
have a higher failure rate when compared to other engineering projects (Berti-Equille and
Borge-Holthoefer, 2015). It is essential to develop the AI-based technologies in a cooperative
manner where there is synergy between IT teams and business units to meet larger goals. It
can be further added that skills and knowledge need to be readjusted for adaptation in a
flexible manner (Leonard-Barton, 1995).
4.3.2 Business process transformation. It has been rightly said that “no business
survives over the long term without reinventing itself”(Bertolini et al.,2015). BPT can be
defined as the transformation of products, processes and organizational aspects. The
genesis of BPT may be in the factors that could potentially make the current processes
inefficient and thereby threaten the sustainability of the firms. The internet changed some of
the business models and slowly brought an end to physical distribution of data.
BPT is essentially a redesign of the business processes with the intention of
improvements in cost, quality, service and speed (Hammer and Champy, 2009). The first
step toward BPT is the same as that of NPD –“identification of opportunities.”This is a
crucial step where the firms may question themselves on certain assumptions and business
processes. One of the questions they can ask is, “how could they serve the customer better
with new technology?”In the case of AI-based solutions like virtual assistants, the firm may
answer the question by saying that it would eliminate the waiting time for customer query,
it would standardize the experience and provide faster resolution of queries.
Similarly, we may ask “how would the technology impact our business?”On similar
lines, it can be answered as –the use of chatbots would reduce the HR requirements and
reduce cost. Another way to question the business model is whether any new technologies
could make the current process simpler. The end goals may be improved customer
VJIKMS
51,3
358
experience, reduced cost of operation, reduced cost of goods sold, improved revenue,
enhanced lead-time, higher quality (Jha et al.,2016). This would then serve as inputs in
deriving the OL and R&D strategies.
Once the firm decides on what processes to improve, it must consider alternatives if
available or benchmark it with existent technology or process to make sure that the
technology is worth upgrading. While this evaluative phase is undertaken, the criteria
should be very clearly laid out. Based on a review of extant literature, Mitropoulos and
Tatum (2000) enumerated five attributes that govern the adoption process:
(1) Compatibility –The “task-technology fit”theory suggests a stronger fit between
the technology and the task for higher performance (Goodhue and Thompson,
1995). The application area of AI must be a holistically seen as an interaction with
the employees, teams, departments of the organization.
(2) Complexity –It is the level of difficulty associated with understanding the
technology like AI that encompasses lots of different types of approaches and
entails varying levels of complexities.
(3) Observability –Unless the organization can contemplate upon and scrutinize the
alternatives they have at hand, they should not move forward.
(4) Triability –Building upon the previous point, the proof of concept (POC) needs to
be tested out. This would reduce the problems later on.
(5) Relative advantage –These POCs are rated and compared based upon their
performance, cost and risk as per the importance of these three attributes.
The comparison helps in selecting the best option from the different concepts. Once the
choice of technology is finalized, the business transformation process can be undertaken.
5. Contemporary implementation themes
5.1 Servicitization
“Servicization is a business strategy to sell the functionality of a product rather than the
product itself”(Örsdemir et al., 2018). The need for establishment and maintenance of
hardware and software is eliminated by a centralized service offering which is charged for
usage and remains on tap. As discussed earlier, in simple terms, it means converting a
product into a service. One of the common servicitization in IT is “cloud computing”where
the computing functions like storage and processing takes place on a remote computer
available to the user as service rather than as a physical product (Mell and Grance, 2011;
Varghese and Buyya, 2018;Belbergui et al.,2017). There is no need to set up, maintain and
upgrade computers, thereby helping in cost cutting. It also increases the accessibility and
use of resources which can be accessed anytime, anywhere.
Cloud computing is usually classified based upon the service that is provided and named
as “X as a service”which is an acronym for “anything as a service.”There are three main
types of XaaS, namely, IaaS –Infrastructure as a Service, PaaS –Platform as a Service and
SaaS –Software as a Service (Kavis, 2014). IaaS is a type of cloud where the cloud vendor
provides only the servers, storage and networking while the client sets up their operating
system and software on the vendor’s servers (IBM.com, 2019). IaaS is most useful for
organizations that do not have physical space or infrastructure. PaaS is a cloud platform
where server as well as operating system and developmental tools are provided by the
vendor. It is useful for rapid and collaborate development of applications. SaaS provides
access to applications through network eliminating the need to install software on client
devices. This eliminates the hassle on maintaining the hardware as well as software.
Business value
359
SaaS is easiest to implement but least flexible, which makes it more suitable for BPT,
while IaaS is more suitable for NPD because of the flexibility and controls it offers at the cost
of reduced ease of use. PaaS is somewhere in between both and could be leveraged for both.
Apart from these, there are some other Xaas implementations that could benefit AI.
Business process management over PaaS (or BPM PaaS) enables execution of customized
business processes on the cloud (Riemann, 2015). These preset instructions can help reduce
time and effort in managing business processes. BPaaS or “business process as a service”is
delivering business process outsourcing (BPO) over cloud (Gartner.com, 2019). Essentially it
is automation of human interactive agents like call center respondents.
There are different applications of AI, right from prediction engines to virtual assistants
to robotics and each case would fit into an appropriate type of an X as a service platform. A
brief overview of three main XaaS architectures: IaaS, PaaS and SaaS and their relevance to
AI is provided in Table I where we present some examples of each of the types of XaaS.
An example of IaaS is Amazon’s Elastic Compute Cloud (EC2) service, which is used by
Airbnb to analyze over 50 GB of data daily (Amazon.com, 2019). EC2 enables load balancing
of huge chunks of static data like user pictures and dynamic data like user activities into
different EC2 instances on cloud. One of the examples of PaaS would be Microsoft’s Azure
ML Service, which offers algorithms for text analytics (Microsoft.com, 2019a). Text
analytics algorithms could be used by a retailer to analyze social media data to ascertain
whether discussions are positive or negative based on sentiment score (Microsoft.com,
2019b). Similarly, a SaaS offered by IBM under Watson Developer Cloud is “Tone
Analyzer,”which can be easily integrated with chatbots and social streams to understand
the emotions and communication styles (IBM.com, 2016).
However, we also need to discuss some of the key drawbacks of cloud-based AI
technologies. Though cloud would provide lower cost, higher quality of service, scalability
and time reduction, it may entail reliability and security issues as well as long-term cost
ineffectiveness. There is a lot of discussion around security and privacy of AI and having
some sensitive data and processing locally. This is possible by using hybrid cloud (Jain and
Hazra, 2019).
5.2 Decentralized value creation
There has been an overall trend toward decentralization of the firm’s role in value creation.
The change from being a “content gatekeeper”to “customer gatekeeper”opens up avenues
in value creation through content creation, infrastructure, access, modules and orchestration
(Pagani, 2013). As per our discussion on the platforms of AI value chain, there can be
different kinds of platforms. The second aspect is the shift of the power from the firms to the
users. This would mean that consumer have more say in what they want, how they want
and when they want. This information may either be explicitly stated by the customer or
Table I.
AI Cloud platforms
Cloud model AI solution Service provider
IaaS Infrastructure like storage, processing and AI engine for
higher customization
Amazon EC2
PaaS Platform to build and deploy AI solutions using reusable
codes and APIs
Amazon Web Services
ML, Microsoft Azure ML
SaaS AI-based apps and software through a subscription model Microsoft Cognitive
Services, Watson
Developer Cloud (IBM)
VJIKMS
51,3
360
implicitly stated in howthey interact with the products. Consecutively, the firms would need
to align to these requirements and develop their strategy from these inputs (Brenner et al.,
2014).
6. Artificial inteligence use cases in two managerial areas
6.1 Enhancing the customer segmentation
One of the traditional but powerful approaches in marketing is customer segmentation.
Segmentation is segregation of different types of buyers who will respond to different kinds
of marketing efforts. Traditionally, it has been successful in enhancing marketing
effectiveness; however, AI can help increase marketing accuracy further by delivering one-
to-one marketing. Micro-segmentation is classification of customers on a finer level and
reveals more nuanced aspects of their preferences, lifestyle and aspirations rather than
broader aspects like price-sensitivity. ML algorithms can be used to map the customer
journey to understand the patterns like effect of change of location on purchase of luxury
products or finding a subset of price-sensitive customers who may purchase specific luxury
products. Micro-segmentation can further enhance the personalization of the marketing
campaigns (Kushmaro, 2018).
An interesting case for micro-segmentation is that of Boeing Employees Credit Union,
which used micro-segmentation to optimize email communications resulting in higher
responses to promotion drives for loans, credit cards and mortgages (Rijn, 2019). This
resulted in a 10 per cent lift for the campaigns.
6.2 Operations and logistic efficiency
There are many applications of AI in the field of operations (Cohen and Sherkat, 1987;Lau
et al., 2009). The use of AI opens up new avenues for managing the operations not just for
firms that indulge in the movement of physical goods but also help improve the service
quality applying the same principles where it is applicable. Overall, the optimization of
operations management (Jacobs et al., 2004) using AI would provide the firm with a
competitive advantage over others because of increased performance because of better
prediction of the volatility of time because of large number of factors. These factors were
difficult to model without using advanced ML techniques.
DHL is one of the leaders in logistic services. They have developedan ML algorithm that
uses over 58 parameters to predict the average transit time a week in advance (Gesing et al.,
2018). This is one of the many use cases that have shown potential of AI in logistics and
supply chain optimization.
7. Discussion and conclusion
AI is an umbrella term for some of the most potent upcoming technologies that could open
up new avenues of development of solutions. It has been projected that AI could analyze the
data that was previously not analyzable, create real-time insights and enhance a firm’s
performance management (Clerck, 2018). Also, it can be applied to automate different kind
of processes, leading to reduced labor requirements and enhanced efficiency. However, a key
difference between successful use of AI and unsuccessful would lie in the acquisition and
development of learning about this technology.
This requires developing technological competencies through OL as well as using the
developed competencies to develop products and services. This bifurcation helps in
simplifying the focus on development or acquisition of scarce resources. Technologies like
chatbots are a great starting point as POC for AI. They are found to be not just a novelty but
offer functional benefits (Shawar and Atwell, 2007). It has already been used in many
Business value
361
organizations worldwide and gains more human-like abilities in expression and problem-
solving, as it learns from theinteractions (Hill et al., 2015). In a way, firms can readily exploit
it by making incremental changes and implanting it in their enterprise. There are also many
enterprises IT solutions making use of AI like the AI-based knowledge mining tools, AI-
based pattern recognition and robotic applications. We have also discussed general
directions of explorative research in marketing as well as operations where AI has been
used. However, as the current applications of AI stand, it is needs breakthroughs to develop
“self-awareness”and reflection to go beyond the automation of tasks and pattern
recognition. Once we develop a way to model such properties, AI can understand the
subjectivity of social milieu to have a collective experiential knowledge (Sanzogni et al.,
2017).
Lastly, there are also concerns regarding ethical and moral use of AI, which we have not
discussed here. We must not forget that the impetus should be on preserving our core
values, our belief systems and sense of well-being for not just humans, but the entire
ecosystem in which we thrive. We need to ensure that AI transformationdoes not happen at
the cost of our ability to think, ability to reflect, ability to take stock and ability to keep the
wheel of world development rotating. It has been said by Elbert Hubbard: “One machine can
do the work of fifty ordinary men. No machine can do the work of one extraordinary man.”
8. Implications, limitations and future directions
8.1 Implications for theory and practice
AI has been under development for over six decades. However, the recent developments
have poised it for becoming the next big disrupter for the businesses (Deloitte, 2019).
Usually, the early adopters gain market share as well as have higher revenues (Dos Santos
and Peffers, 1995;O’Connor et al., 1998). We have discussed the approaches toward
developing competencies in AI through OL. Though the goals of a theorist and practitioner
may be different, a good theoretical model could serve both of them well (Dubin, 1976).
This paper has several implications from a manager’s perspective. Firstly, the roadmap
should be useful for the managers who are not sure where to begin. The overview of current
AI and OL literature could inform the strategic directions for AI adoption. Secondly, we
have shown that knowledge management (KM) should be the starting point while the
managers need to consider product or process innovation as mediator between KM and firm
performance (Turulja and Bajgori
c, 2018). Thirdly, the two broad approaches to innovation
through NPD and BPT could help in defining organizational priorities and help in extracting
firm performance through innovation. The proposed roadmap is flexible enough to cater to
different organizations and could serve as a guiding light in navigating the uncertain
landscape of AI adoption and value appropriation.
8.2 Limitations and future directions
Through this paper, we have only touched theoretical foundations of rich literature in
strategy and IS. This rudimentary overview of concepts need much deeper understanding
and analysis and a deep dive could inform the reader about them in detail. This discussion
aimed to provide an executive’s overview of the AI strategy. However, internal technology
adoption is challenging. At an operational level, there are many factors that govern the
technology adoption and continued usage behaviors (Karahanna et al., 1999), but we
have not discussed about the adoption, rather limiting this discussion to learning and
R&D. The discussion presented here has been developed from extant literature.
However, it could be further developed through an empirical inquiry. Future work can
look into this area. Ambidexterity is one of the most difficult tradeoffs for an
VJIKMS
51,3
362
organization (O’Reilly and Tushman, 2004), and the simplistic presentation of the
roadmap should not be taken as a remedy pill and needs due diligence. One of the major
challenges in development would be to conceptualize and create AI that can have
cognitive awareness, which is best described as “subjective awareness”of the world
(Nagel, 1974). Future researchers can provide basis for such holistic approaches that go
beyond automation of tasks. There are many challenges in developing AI as well as
implementing those developments. This presents many opportunities too for researchers
as well as practitioners and they can look into these as future directions.
References
Agrawal, A., Gans, J.S., and Goldfarb, A. (2017), What to Expect from Artificial Intelligence, MIT Sloan
Management Review.
Amazon.com (2019), “Airbnb case study –amazon web services (AWS)”, Amazon Web Services,
available at: https://aws.amazon.com/solutions/case-studies/airbnb/ (accessed 25 May 2019).
Bahdanau, D. Cho, K. and Bengio, Y. (2014), “Neural machine translation by jointly learning to align
and translate”, ArXiv Preprint ArXiv:1409.0473.
Baryannis, G., Validi, S., Dani, S. and Antoniou, G. (2019), “Supply chain risk management and artificial
intelligence: state of the art and future research directions”,International Journal of Production
Research, Vol. 57 No. 7, pp. 2179-2202.
Belbergui, C., Elkamoun, N. and Hilal, R. (2017), “Cloud computing: overview and risk identification
based on classification by type”,2017 3rd International Conference of Cloud Computing
Technologies and Applications (CloudTech), presented at the 2017 3rd International Conference
of Cloud Computing Technologies and Applications (CloudTech), pp. 1-8.
Berti-Equille, L. and Borge-Holthoefer, J. (2015), “Veracity of data: from truth discovery computation
algorithms to models of misinformation dynamics”,Synthesis Lectures on Data Management,
Vol. 7 No. 3, pp. 1-155.
Bertolini, M., Duncan, D. and Waldeck, A. (2015), “Knowing when to reinvent”,Harvard Business
Review, Vol. 93 No. 12, pp. 90-101.
Brenner, W., Karagiannis, D., Kolbe, L., Krüger, J., Leifer, L., Lamberti, H.-J., Leimeister, J.M., Österle, H.,
Petrie, C., Plattner, H. and Schwabe, G. (2014), “User, use and utility research: the digital user as
new design perspective in business and information systems engineering”,Business and
Information Systems Engineering, Vol. 6 No. 1, pp. 55-61.
Brynjolfsson, E. and McAfee, A. (2016), The Second Machine Age: Work, Progress, and Prosperity in a
Time of Brilliant Technologies, 1st ed., W. W. Norton and Company, New York, NY, London.
Bushnel, M. (2018), “AI Faceoff: Siri vs Cortana vs Google assistant vs Alexa”, available at: www.
businessnewsdaily.com/10315-siri-cortana-google-assistant-amazon-alexa-face-off.html (accessed
7 January 2019).
Carlile, P.R. (2002), “A pragmatic view of knowledge and boundaries: boundary objects in new product
development”,Organization Science, Vol. 13 No. 4, pp. 442-455.
Chakravarty, A., Grewal, R. and Sambamurthy, V. (2013), “Information technology competencies,
organizational agility, and firm performance: enabling and facilitating roles”,Information
Systems Research, Vol. 24 No. 4, pp. 976-997.
Cheon, M.J., Grover, V. and Teng, J.T.C. (1995), “Theoretical perspectives on the outsourcing of
information systems”,Journal of Information Technology, p. 11.
Chesbrough, H.W. (2006), Open Innovation: The New Imperative for Creating and Profiting from
Technology, Harvard Business Press.
Choi, J.J., Ju, M., Kotabe, M., Trigeorgis, L. and Zhang, X.T. (2018), “Flexibility as firm value driver:
evidence from offshore outsourcing”,Global Strategy Journal, Vol. 8 No. 2, pp. 351-376.
Business value
363
Clerck, J.-P.D. (2018), “Corporate performance management –role of AI and new technologies”,I-
SCOOP, 16 April, available at: www.i-scoop.eu/corporate-performance-management-ai-new-
technologies/ (accessed 12 April 2019).
Cohen, A.I. and Sherkat, V.R. (1987), “Optimization-based methods for operations scheduling”,
Proceedings of the IEEE, Vol. 75 No. 12, pp. 1574-1591.
Crossan, M.M. and Berdrow, I. (2003), “Organizational learning and strategic renewal”,Strategic
Management Journal, Vol. 24 No. 11, p. 1087.
Davis, G. (2015), “Poor research-industry collaboration: time for blame or economic reality at work?”,
The Conversation, available at: http://theconversation.com/poor-research-industry-collaboration-
time-for-blame-or-economic-reality-at-work-50306 (accessed 30 November 2019).
Deloitte (2019), “Artificial intelligence disruption”, Deloitte United States, available at: www2.deloitte.
com/us/en/pages/technology-media-and-telecommunications/articles/artificial-intelligence-
disruption.html (accessed 22 April 2019).
DeSanctis, G., Glass, J.T. and Ensing, I.M. (2002), “Organizational designs for R&D”,Academy of
Management Perspectives, Vol. 16 No. 3, pp. 55-66.
Dirican, C. (2015), “The impacts of robotics, artificial intelligence on business and economics”,Procedia
- Social and Behavioral Sciences, Vol. 195, pp. 564-573.
Dos Santos, B.L. and Peffers, K. (1995), “Rewards to investors in innovative information technology
applications: first movers and early followers in ATMs”,Organization Science, Vol. 6 No. 3,
pp. 241-259.
Dougal, J. (2018), “Banks excited by AI, but uncertainty remains”, NCR, 19 April, available at: https://
10.43.32.55/content/ncr/us/en/home/company/blogs/financial/banks-excited-by-ai-but-uncertainty-
remains (accessed 23 May 2019).
Dubin, R. (1976), “Theory building in applied area”,Handbook of Industrial and Organizational
Psychology, R& McNally College.
Duff, A.S. (2004), “The past, present, and future of information policy: towards a normative
theory of the information, society”,Information, Communication and Society,Vol.7No.1,
pp. 69-87.
Enkel, E., Gassmann, O. and Chesbrough, H. (2009), “Open R&D and open innovation: exploring the
phenomenon”,R&D Management, Vol. 39 No. 4, pp. 311-316.
Fiol, C.M. and Lyles, M.A. (1985), “Organizational learning”,Academy of Management Review, Vol. 10
No. 4, pp. 803-813.
Gartner.com (2019), “Business process as a service (bpaas)”, Gartner, available at: www.gartner.com/
en/information-technology/glossary/business-process-as-a-service-bpaas (accessed 10 December
2019).
Gesing, B., Peterson, S.J., and Michelsen, D. (2018), Artificial Intelligence in Logistics, Whitepaper,DHL
Customer Solutions andInnovation, Troisdorf, p. 45.
Goodhue, D.L. and Thompson, R.L. (1995), “Task-technology fit and individual performance”,MIS
Quarterly, Vol. 19 No. 2, pp.213-236.
Gregory, M.J. (1995), “Technology management: a process approach”,Proceedings of the
Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture,Vol.209
No. 5, pp. 347-356.
Hall, D.W. and Pesenti, J. (2017), “Growing the artificial intelligence industry in the UK”, Independent
Review for the Department for Digital, Culture, Media and Sport/Department for Business,
Energy and Industrial Strategy, available at: www.Gov.Uk/Government/Publications/Growing-
the-Artificial-Intelligence-Industry-in-the-Uk
Hammer, M. and Champy, J. (2009), Reengineering the Corporation: Manifesto for Business Revolution,
A, Zondervan.
VJIKMS
51,3
364
Hao, K. (2018), “What is machine learning?”, MIT Technology Review, available at: www.
technologyreview.com/s/612437/what-is-machine-learning-we-drew-you-another-flowchart/
(accessed 30 November 2019).
Hill, S.A. and Birkinshaw, J. (2014), “Ambidexterity and survival in corporate venture units”,Journal of
Management, Vol. 40 No. 7, pp. 1899-1931.
Hill, J., Ford, W.R. and Farreras, I.G. (2015), “Real conversations with artificial intelligence: a
comparison between human–human online conversations and human–chatbot conversations”,
Computers in Human Behavior, Vol. 49, pp. 245-250.
IBM.com (2016), “Watson tone analyzer”, 28 November, available at: www.ibm.com/watson/services/
tone-analyzer/ (accessed 25 May 2019).
IBM.com (2019), “IaaS PaaS SaaS cloud service models”, IBM, Corporate, 20 November, available at:
www.ibm.com/in-en/cloud/learn/iaas-paas-saas (accessed 10 December 2019).
Jacobs, F.R., Chase, R.B. and Aquilano, N. (2004), Operations Management for Competitive Advantage,
Vol. 64, McGraw Hill, Boston, p. 70.
Jain, T. and Hazra, J. (2019), “Hybrid cloud computing investment strategies”,Production and
Operations Management, Vol. 28 No. 5, doi: 10.1111/poms.12991.
Jang, S., Hong, K., Woo Bock, G. and Kim, I. (2002), “Knowledge management and process innovation:
the knowledge transformation path in Samsung SDI”,Journal of Knowledge Management, Vol. 6
No. 5, pp. 479-485.
Jha, M., Jha, S. and O’Brien, L. (2016), “Combining big data analytics with business process using
reengineering”,2016 IEEE Tenth International Conference on Research Challenges in
Information Science (RCIS), IEEE, pp. 1-6.
Johnson, M., Schuster, M., Le, Q.V., Krikun, M., Wu, Y., Chen, Z., Thorat, N., Viégas, F., Wattenberg, M.,
Corrado, G. and Hughes, M. (2017), “Google’s multilingual neural machine translation system:
enabling zero-shot translation”,Transactions of the Association for Computational Linguistics,
Vol. 5, pp. 339-351.
Karahanna, E., Straub, D.W. and Chervany, N.L. (1999), “Information technology adoption across time:
a cross-sectional comparison of pre-adoption and post-adoption beliefs”,MIS Quarterly, Vol. 23
No. 2, pp. 183-213.
Kavis, M.J. (2014), Architecting the Cloud: Design Decisions for Cloud Computing Service Models (SaaS,
PaaS, and IaaS), John Wiley and Sons.
Kuittinen, H., Puumalainen, K., Jantunen, A., Kyläheiko, K. and Pätäri, S. (2013), “Coping with
uncertainty –exploration, exploitation, and collaboration in R&D”,International Journal of
Business Innovation and Research, Vol. 7 No. 3, p. 340.
Kurzweil, R. (2006), The Singularity is near: When Humans Transcend Biology, Penguin USA, New
York, NY.
Kurzweil, R., Richter, R., Kurzweil, R., and Schneider, M.L. (1990), The Age of Intelligent Machines,
Vol. 579, MIT pressCambridge, MA.
Kushmaro, P. (2018), “How AI is reshaping marketing”, CIO, 4 September, available at: www.cio.com/
article/3302739/how-ai-is-reshaping-marketing.html (accessed 25 May 2019).
Lau, H.C., Chan, T.M., Tsui, W.T., Ho, G.T. and Choy, K.L. (2009), “An AI approach for optimizing
multi-pallet loading operations”,Expert Systems with Applications, Vol. 36 No. 3, pp. 4296-4312.
LeCun, Y., Bengio, Y. and Hinton, G. (2015), “Deep learning”,Nature, Vol. 521 No. 7553, p. 436.
Leonard-Barton, D. (1995), “Wellsprings of knowledge: building and sustaining the sources of innovation”,
SSRN Scholarly Paper No. ID 1496178, Social Science Research Network, Rochester, New York, NY,
available at: https://papers.ssrn.com/abstract=1496178 (accessed 11 January 2019).
Levitt, B. and March, J.G. (1988), “Organizational learning”,Annual Review of Sociology, Vol. 14 No. 1,
pp. 319-338.
Business value
365
Liao, S.-H., Wu, C.-C., Hu, D.-C. and Tsuei, G.A. (2009), “Knowledge acquisition, absorptive capacity,
and innovation capability: an empirical study of Taiwan’s knowledge-intensive industries”,
Technology, Vol. 11, p. 13.
Lichtenthaler, U. (2011), “Open innovation: past research, current debates, and future directions”,
Academy of Management Perspectives, Vol. 25 No. 1, pp. 75-93.
Liu, W. (2006), “Knowledge exploitation, knowledge exploration, and competency trap”,Knowledge and
Process Management, Vol. 13 No. 3, pp. 144-161.
Loh, L. and Venkatraman, N. (1995), “An empirical study of information technology outsourcing:
benefits, risks, and performance implications”, p. 13.
Mckinsey.com (2018), “Embrace the uncertainty of AI”, 23 July, available at: www.mckinsey.com/
business-functions/organization/our-insights/the-organization-blog/embrace-the-uncertainty-of-
ai (accessed 23 May 2019).
March, J.G. (1991), “Exploration and exploitation in organizational learning”,Organization Science, Vol.
2 No. 1, pp. 71-87.
Mell, P. and Grance, T. (2011), The NIST Definition of Cloud Computing, Computer Security Division,
Information Technology Laboratory, National.
Metz, C. (2017), “Tech giants are paying huge salaries for scarce A.I. Talent”, The New York Times, 22
October, available at: www.nytimes.com/2017/10/22/technology/artificial-intelligence-experts-
salaries.html (accessed 9 May 2019).
Microsoft.com (2019a), “Text analytics –azure machine learning studio”, 6 May, available at: https://
docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/text-analytics
(accessed 25 May 2019).
Microsoft.com (2019b), “Perform sentiment analysis with text analytics REST API –azure cognitive
services”, Microsoft Azure Cognitive Services, 21 November, available at: https://docs.microsoft.
com/en-us/azure/cognitive-services/text-analytics/how-tos/text-analytics-how-to-sentiment-
analysis (accessed 12 December 2019).
Mitropoulos, P. and Tatum, C.B. (2000), “Forces driving adoption of new information
technologies”,Journal of Construction Engineering and Management,Vol.126No.5,
pp. 340-348.
Nagel, T. (1974), “What is it like to be a bat?”,The Philosophical Review, Vol. 83 No. 4,pp. 435-450.
Nieto, M. (2004), “Basic propositions for the study of the technological innovation process in the firm”,
European Journal of Innovation Management, Vol. 7 No. 4, pp. 314-324.
Norman, D.A. and Verganti, R. (2014), “Incremental and radical innovation: design research vs
technology and meaning change”,Design Issues, Vol. 30 No. 1, pp. 78-96.
Nott, G. (2018), “Don’t worry, world-leading AI expert doesn’t know what AI is either”, Computerworld,
available at: www.computerworld.com.au/article/644740/don-t-worry-world-leading-ai-expert-
doesn-t-know-what-ai-either/ (accessed 9 May 2019).
O’Connor, G., Kung, H. and O’Keefe, R.M. (1998), “Early adopters of the web as a retail medium: small
company winners and losers”,European Journal of Marketing, Vol. 32 Nos 7/8, pp. 629-643.
O’Reilly, C.A., 3rd. and Tushman, M.L. (2004), “The ambidextrous organization”,Harvard Business
Review, Vol. 82 No. 4, p. 74.
OECD (2019), University-Industry Collaboration: New Evidence and Policy Options, OECD Publishing,
Paris, p. 120.
Örsdemir, A., Deshpande, V. and Parlaktürk, A.K. (2018), “Is servicization a win-win strategy?
Profitability and environmental implications of servicization”,Manufacturing and Service
Operations Management, doi: 10.1287/msom.2018.0718.
Pagani, M. (2013), “Digital business strategy and value creation: framing the dynamic cycle of control
points”,MIS Quarterly, Vol. 37 No. 2, pp. 617-632.
VJIKMS
51,3
366
Paunov, C., Borowiecki, M. and El-Mallakh, N. (2019), “Cross-country evidence on the contributions of
research institutions to innovation”,OECD Science, Technology and Industry Policy Papers,
doi: 10.1787/d52d6176-en.
Ramiller, N.C., Davidson, E., Wagner, E.L., and Sawyer, S. (2008), “Turning products into services and
services into products: contradictory implications of information technology in the service
economy”, in Barrett, M., Davidson, E., Middleton, C. and DeGross, J.I. (Eds), Information
Technology in the Service Economy: Challenges and Possibilities for the 21st Century, Springer,
pp. 343-348.
Ransbotham, S., Kiron, D., Gerbert, P. and Reeves, M. (2017), “Reshaping business with artificial
intelligence: closing the gap between ambition and action”,MIT Sloan Management Review,
Vol. 59 No. 1.
Real, J.C., Leal, A. and Roldán, J.L. (2006), “Information technology as a determinant of organizational
learning and technological distinctive competencies”,Industrial Marketing Management, Vol. 35
No. 4, pp. 505-521.
Reitman, W.R. (1984), Artificial Intelligence Applications for Business: Proceedings of the NYU
Symposium, May, 1983, Intellect Books.
Riemann, U. (2015), “Benefits and challenges for BPM in the cloud”,International Journal of
Organizational and Collective Intelligence, Vol. 5 No. 1, pp. 32-61.
Rijn, J.V. (2019), “AI finally makes micro segmentation a reality for financial marketers”, The Financial
Brand, 22 October, available at: https://thefinancialbrand.com/89462/ai-micro-segmentation-
financial-institution-marketers/ (accessed 21 January 2020).
Rodríguez, P., Kuvaja, P., and Oivo, M. (2014), “Lessons learned on applying design science for bridging
the collaboration gap between industry and academia in empirical software engineering”,
Proceedings of the 2nd International Workshop on Conducting Empirical Studies in Industry,
ACM, pp. 9-14.
Roos, J.G. and Kern, C.F. (1996), Modelling Customer Demand Response to Dynamic Price Signals Using
Artificial Intelligence, IET.
Sambamurthy, V., Bharadwaj, A. and Grover, V. (2003), “Shaping agility through digital options:
reconceptualizing the role of information technology in contemporary firms”,MIS
Quarterly, pp. 237-263.
Sanzogni, L., Guzman, G. and Busch, P. (2017), “Artificial intelligence and knowledge management:
questioning the tacit dimension”,Prometheus, Vol. 35 No. 1, pp. 37-56.
Shawar, B.A. and Atwell, E. (2007), “Chatbots: are they really useful?”,Ldv Forum, Vol. 22,pp. 29-49.
Sivathanu, B. and Pillai, R. (2018), “Smart HR 4.0 –how industry 4.0 is disrupting HR”,Human
Resource Management International Digest, Vol. 26 No. 4, pp. 7-11.
Sterne, J. (2017), Artificial Intelligence for Marketing: Practical Applications, John Wiley and Sons.
Stone, C.B., Neely, A.R., and Lengnick-Hall, M.L. (2018), “Human resource management in the digital
age: big data, HR analytics and artificial intelligence”,Management and Technological
Challenges in the Digital Age, CRC Press, pp. 13-42.
Teece, D.J. (1998), “Capturing value from knowledge assets: the new economy, markets for know-how,
and intangible assets”,California Management Review, Vol. 40 No. 3, pp. 55-79.
Tirgul, C.S. and Naik, M.R. (2016), “Artificial intelligence and robotics”,International Journal of
Advanced Research in Computer Engineering and Technology, Vol. 5 No. 6, pp. 1787-1793.
Tsoukas, H. and Vladimirou, E. (2001), “What is organizational knowledge?”,Journal of Management
Studies, Vol. 38 No. 7, pp. 973-993.
Turulja, L. and Bajgori
c, N. (2018), “Knowledge acquisition, knowledge application, and innovation
towards the ability to adapt to change”,International Journal of Knowledge Management,
Vol. 14 No. 2, pp. 1-15.
Business value
367
Tushman, M.L. and O’Reilly, C.A. III, (1996), “Ambidextrous organizations: managing evolutionary and
revolutionary change”,California Management Review, Vol. 38 No. 4, pp. 8-29.
Varghese, B. and Buyya, R. (2018), “Next generation cloud computing: new trends and research
directions”,Future Generation Computer Systems, Vol. 79, pp. 849-861.
Watson, H.J. (2017), “Preparing for the cognitive generation of decision support”,MIS Quarterly
Executive.
Webster, F. (2007), Theories of the Information Society, Taylor and Francis, Hoboken, available at:
http://public.eblib.com/choice/publicfullrecord.aspx?p=292956 (accessed 7 May 2019).
Wu, Y. Schuster, M. Chen, Z. Le, Q.V. Norouzi, M. Macherey, W. Krikun, M. Cao, Y. Gao, Q. Macherey,
K. and Klingner, J. (2016), “Google’s neural machine translation system: bridging the gap
between human and machine translation”, ArXiv Preprint ArXiv:1609.08144.
Corresponding author
Arindra Nath Mishra can be contacted at: arindra@astra.xlri.ac.in
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
VJIKMS
51,3
368
