Journal of Information Technology Case and Application

Research

ISSN: 1522-8053 (Print) 2333-6897 (Online) Journal homepage: https://www.tandfonline.com/loi/utca20

The AI advantage: how to put the artificial

intelligence revolution to work
by Thomas H. Davenport, Cambridge, MA, MIT Press, 2018, 248 pp., $19.95
(paper back), $9.20 (eText book), ISBN 9780262538008

Vijay K. Vemuri

To cite this article: Vijay K. Vemuri (2020): The AI advantage: how to put the artificial intelligence
revolution to work, Journal of Information Technology Case and Application Research, DOI:
10.1080/15228053.2020.1756084

To link to this article: https://doi.org/10.1080/15228053.2020.1756084

Published online: 22 Apr 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=utca20
JOURNAL OF INFORMATION TECHNOLOGY CASE AND APPLICATION RESEARCH

BOOK REVIEW

The AI advantage: how to put the artificial intelligence revolution to work, by Thomas
H. Davenport, Cambridge, MA, MIT Press, 2018, 248 pp., $19.95 (paper back), $9.20 (eText
book), ISBN 9780262538008

JITCAR’s review of books on the role of artificial intelligence (AI) started with a review of Prediction
Intelligence (Heath, 2019). Continuing with this theme, we now review The AI Advantage by
Thomas H. Davenport (MIT Press 2018).
This book should be of interest to many, especially to senior managers and academics. Very few
operations of any business are untouched by advances in AI. The question is no longer, “do we need
AI for our organization,” but how our organization can incorporate AI in the overall strategy in the
most effective way. The AI Advantage is not about AI hype, or a list of “gee whiz” AI applications,
but how businesses should adapt to AI and leverage their capabilities. The author’s frequent use of
“cognitive technologies” to refer to AI (these two terms are used interchangeably in the book, and in
this review) is fitting in that it takes away the unnecessary preconceived biases, or lofty expectations
a reader may have of AI. The AI advantage emphasizes how it can transform organizations, as well
as, how AI provides a road map to harness an organization’s full potential. The book also articulates
implementation strategies and roll out possibilities. The author utilizes his extensive experience as
a consultant to AI projects and his access to senior management to document strategies for AI
implementation. The book supports its arguments by referring to several surveys conducted by large
consulting and technology firms, notably Deloitte’s 2017 “cognitive aware” survey of senior man-
agers. The primary view point is that cognitive technologies are mainly about gaining insights from
data, information, and knowledge. This perspective is refreshingly different from many who hold the
narrow view that AI will eventually create robots to replace humans in performing various tasks. In
what follows, we will refer to the book as AI Advantage at many points.
The book is divided into eight chapters, each chapter focusing on the role of AI in business
processes, jobs and skills, and social & ethical issues. Chapter 1 provides several cases of AI
deployment in organizations, highlighting experiences of the Oncology Expert Adviser program
within MD Anderson Cancer Centers. The tone of the chapter is not one of contributing to the AI
hype, but lowering AI expectations by providing a balanced view. It suggests that in the short run AI
will have evolutionary benefits, but in the long run it is likely to be revolutionary. A taxonomy
(adopted from other sources, which the author calls “periodic table”) of AI and their applications is
provided and further discussed. The AI technologies are classified into 1) statistical and machine
learning, 2) neural networks, 3) deep learning, 4) natural language programming, 5) rules-based
expert systems, 6) physical robots, and 7) robotics process automation.
Opportunities and challenges of AI in the enterprises are explained in Chapter 2. Most of the
conclusions of the chapter are supported by the 2017 Deloitte’s cognitive awareness survey of 250
managers of organizations that are already pursuing AI projects. Well, over 75% of the respondents
believe cognitive technologies are important for improving products, services, and business processes
and their companies will be substantially transformed within the next 3 years (that is by 2020). In
addition to the Deloitte survey, AI advantage draws on results of recent AI surveys by McKinsey,
CIO, and Teradata. Details of cognitive technology deployment in health care, marketing, and
manufacturing industries are explained in detail.
Chapter 3 summarizes characteristics of 152 projects the author and Deloitte Consulting parti-
cipated in. Although the sample is not random, results show that health care, financial services,
consumer product, and industrial product industries are involved in nearly two-thirds of the AI
projects. AI involvement is not just limited to private sector. Public sector, including US federal
2 BOOK REVIEW

government, comprises 10% of Deloitte’s clients. Among these projects, client motivations are
mainly in 1) automating repetitive processes, 2) gaining insight through analysis, and 3) engaging
with customers and employees. One surprising observation from the majority of these projects is
that staff reduction was neither a primary objective, nor an outcome. It is recognized that early
successes of deploying cognitive technologies are modest, but the author expects that information
intensive sectors such as marketing, health care, financial services can become more valuable and less
expensive due to the use of cognitive technologies.
In Chapter 4, strategies for an organization to benefit from cognitive technologies are discussed.
Monetizing content, acquisition of talent, partnering in building projects, and the extent of deploy-
ment of AI are stressed as important strategies. This chapter discusses many strategic paths that are
available and then how to choose from these strategies. An example of the author’s recommenda-
tions with respect to extent of deployment is to start with a series of smaller, and less ambitious
projects in the area of businesses that the organization has extensive knowledge in. He further
advises the management not to succumb to pressure from AI vendors and consultants to undertake
many large projects all at one time.
Chapter 5 can be called the main chapter of the book. What is AI and what capabilities does it
encompass is a frequent source of disagreement. The definition of AI is not precise, and it can mean
many things to many people. Grouping many different technologies and capabilities into a broad
category of AI creates confusion, and leads to misunderstanding of the AI’s capabilities, and as to
how these can be deployed in organizations. A further discussion of the AI technology classifications
provided in Chapter 1 will clarify any misconceptions in equating AI with robotics. In Chapter
8 cognitive technology tasks, and business processes that can be implemented today are identified.
These should provide senior management assurance that many applications of AI are ready for
implementation now. The tasks that cognitive technologies are already capable to do today are: 1)
implementing detailed predictions and classification models, 2) performing structured tasks, 3)
manipulating large amount of information, 4) understanding human speech and text, 5) planning
and optimizing operations, 6) recognizing images, 7) moving autonomously, and 8) understanding
human emotions. For each of these tasks, process applications, organizational implications, and
challenges to implementation are presented.
The prevalent fear of job losses and new skills requirements in an automated work environment is
the subject of Chapter 6. The book cites study by Frey and Osborne (2017) which uses occupation-
based analysis to conclude that nearly 47% of the existing job classifications can be fully automated.
The earlier drafts of this study generated widespread fear of automation and negative popular view of
AI. The AI Advantage cites several other recent studies predicting a more balanced view of
automation. In this balanced view, AI augments the productivity of human workers. A strong
argument is made that availability of technology to replace the human workers is not the only
factor in automation decision. They cite 1) cost to automate, 2) the wage rates of workers performing
the tasks to be automated, 3) quality improvements due to automation, and 4) regulatory and social
acceptance. Emphasizing the likely near-term scenario of AI augmenting human workers and not
replacing them, the book predicts employment reduction of only 5%, rather than the bleak predic-
tion of 47% mentioned above. A persuasive argument that AI replaces tasks, but not ensemble of all
tasks performed by a single worker, is made. The employment patterns of bank tellers over the past
decades reveal that in the short run bank ATM’s and internet banking complement human tellers
and create higher demand for banking services. The chapter provides sobering predictions of future
jobs based on several studies and critical analysis.
Development and implementation of AI technology are the subject matter of Chapter 7.
Davenport argues that the current operational knowledge of a business enables it to identify
opportunities for implementing cognitive technologies. The road map to implement cognitive
technologies include 1) building own applications, 2) managing collection of disparate cognitive
tools instead of one single integrated cognitive technology platform, and 3) developing applications
with open-source tools. Based on cumulative knowledge gained from wide-spread use of robotics in
 JOURNAL OF INFORMATION TECHNOLOGY CASE AND APPLICATION RESEARCH 3

manufacturing, robotic process automation (RPA) is recommended as an entry-level strategy for

implementing AI. He reaches this conclusion by arguing that RPA is not disruptive, as it does not
change underlying systems it connects to, leading to manageable first-level applications of AI.
Another suggestion is to build broad cognitive platform with vendor (such as IBM’s Watson)
supplied platform. Details of implementation of cognitive technologies at Bank of Montreal, and
GlaxoSmithKline and their outcomes provide a preview of tangible benefits that can be attained.
The final chapter (Chapter 8) is a brief guide to managing the organizational, social, and ethical
implementation of AI. The author takes a balanced approach of recognizing that AI can be
a solution, as well as, a problem in many social interactions. The inability of Facebook’s AI
algorithms to recognize fake news, and spread of racial hatred is a case in point. Although it is
not a problem of cognitive technologies, several examples of algorithmic bias leading to social harm
are cited. Interestingly, and in contrast to dire predictions of job losses due to AI, Facebook is adding
about 10,000 humans in supplementing its algorithms to fact check and to detect potential fake, or
controversial news and posts. Algorithmic bias in existing cognitive technologies used in credit
granting and criminal justice system is also cited as a significant issue.
The AI technologies have matured enough for implementation today. However, senior manage-
ment needs to have a clear understanding of capabilities, benefits, potential pitfalls, and road map for
deployment of AI applications. The AI Advantage is a valuable contribution to understanding as
how to effectively blend AI and business strategy. The author with his extensive experience in
developing and implementing AI projects takes an appropriately balanced view of AI by pointing
potential short comings of AI. Most of the conclusions and recommendations are supported by
responses of several surveys of senior management. These are appropriately referenced and exten-
sively discussed. The recommendation of starting slow by taking up modest applications using
RPA is particularly important. AI Advantage is a must read for senior management of medium to
large organizations. Further, it is an essential reading for academics with interests in information
technology strategies. It can be an excellent additional reading for several graduate information
systems classes.

Vijay K. Vemuri
Quinnipiac University
[email protected]

© 2020 Vijay K. Vemuri

https://doi.org/10.1080/15228053.2020.1756084

Notes on contributor
Vijay K. Vemuri is a Visiting Professor of Accounting at Quinnipiac University. He received his Ph.D. and M.
A.S. from the University of Illinois at Urbana-Champaign. His primary research interests are economic
consequences of information technology and accounting information systems. He has published extensively
in several academic journals.

References
Frey, C. A., & Osborne, M. A. (2017). The future of employment: How susceptible are jobs to computerisation?
Technological Forecasting and Social Change, 114, 254–280. doi:10.1016/j.techfore.2016.08.019
Heath, D. R. (2019). Review of “Prediction machines: The simple economics of artificial intelligence”. Journal of
Information Technology Case and Application Research, 22(19), 163-166.