Making AI Responsible and Effective

Digital Business
Making AI Responsible
– And Effective
To bring AI into the business mainstream, companies need to
complement their technology advances with a focus
on governance that drives ethics and trust. If they don’t,
their AI efforts will fall short of their expectations and
lag the business results delivered by competitors that
responsibly embrace machine intelligence, our latest research
findings suggest.
October 2018

2 / Making AI Responsible – And Effective
Digital Business
Executive Summary
Artificial intelligence (AI) has migrated from science fiction to lab experiments and, now, to
actual business operations. Most people already encounter these technologies as they read
product recommendations on Amazon, receive automated fraud alerts from credit card
companies or ask their Google Assistant to play a certain song. But AI is also at work in less
visible ways, as well – scoring bank customers for creditworthiness, analyzing warranty claims
to uncover upstream production problems, grading college essays and even helping courts
determine how to sentence criminals.
Nevertheless, businesses are in the early stages of AI adoption, and companies are still
learning how to put the technology to work. The challenge today is less about understanding
technical questions and technology capabilities, and more about crafting a strategy,
determining the governance structures and practices needed for “responsible AI,” and
accelerating the move from experiments to full-scale AI adoption.
Earlier this year, we assessed the state of AI in business through the eyes of 975 executives
across industries in the U.S. and Europe (see Methodology, page 22), asking them about their
attitudes, expectations and plans. Key findings include:
❙ Executives are enthusiastic about the importance and potential benefits of AI.
Roughly two-thirds of executives say that AI is extremely or very important to their
companies today, and more than eight out of 10 expect this will be the case three years
from now. About three-quarters foresee major or significant benefits from the technology,
in terms of lower costs; increased revenues; and the ability to introduce new products or
services, or to enter new businesses.
❙ Faster-growing companies are more optimistic and aggressive when it comes to
AI. Executives at faster-growing companies are more likely than others to view AI as
important, and more likely to expect major benefits in the coming years. Eighty-six percent
of executives at these companies say AI is extremely or very important to their company’s
success, compared with 57% of those at their slower-growing competitors. These industry
leaders say they plan to use AI to drive further growth, solidifying their leading positions
and pulling even further away from the pack.
❙ There is a major disconnect between executives’ optimism about AI and actual AI
implementation. While two-thirds of executives said they knew about an AI project
at their company, only 24% of that group – just 15% of all respondents – were aware of
projects that were fully implemented., with the large majority of AI projects still in the
early planning or pilot stages. This contrasts sharply with the fact that 84% of respondents
believe AI will be critically important three years from now.
❙ Many companies appear to lack a strategic focus for AI. When identifying the
challenges they face in implementing AI and the technologies they’re targeting, executive
responses were relatively undifferentiated in both areas – essentially adding up to “all
of the above.” This suggests that companies have yet to hone their strategic plans for

Making AI Responsible – And Effective / 3
Digital Business
integrating AI into the business core. In addition, roughly 40% of respondents said that securing
senior management commitment, buy-in by the business and even adequate budget were
extremely or very challenging, indicating that many companies are not yet fully committed to
AI’s central role in advancing business objectives.
❙ Many companies are not addressing the ethical dimensions of AI. Only about half of the
executives surveyed said their companies have policies and procedures in place to identify
and address ethical considerations – either in the initial design of AI applications or in their
behavior after the system is launched. We believe organizations need to focus on an ethics
lens upfront in the strategy/value stream when deciding whether a problem/need has to be
solved through AI.
What’s more, about two-thirds of that group said the ethical governance structures they
have in place are extremely or very effective. That view may be overly optimistic, and an
indication that many executives are underestimating or unfamiliar with the challenging
ethical questions likely to emerge as AI becomes more sophisticated and pervasive both in its
development and use.
Companies will need to take action on several fronts if they are to achieve significant business
benefits with AI. The key agenda items include:
❙ Formulate AI strategies. These should focus on opportunities that promise measurable
value – not only reduced costs and increased revenue, but also benefits such as improved
customer service, entry to new lines of business and enhanced employee experiences. It’s
especially critical that strategies take a human-centric view of AI, so that machines can work
successfully alongside and for people.
❙ Develop governance structures. Companies will need to work proactively to ensure that
AI decision-making is transparent to those involved; that AI earns trust by avoiding errors
and data-driven biases; and that AI is personalized and able to provide tailored, relevant and
context-aware support as it interacts with humans.
❙ Create and maintain responsible AI applications. Because of AI’s potential ubiquity and
power, ethical concerns need to be interwoven into everything companies do with the
technology. That means building AI systems ethically, and then providing oversight to ensure
that those systems operate ethically over time, even as the AI applications learn and evolve.
To be successful, companies will need to boost their ethics-related efforts upfront as AI
touches more and more parts of business and society.
In many ways, these non-technical considerations – regarding trust, transparency, ethics and
human-centric approaches – are more critical and complex than those related to developing and
running the technology itself. But they are absolutely critical to the success of AI, because they will
determine how well these technologies can be leveraged to advance business objectives.
AI operates in the real world. Companies that do not find solutions in these areas could see
their AI efforts fall short, or fail completely. This, in turn, has the potential to irritate customers,
alienate employees, drive up RD and deployment costs, and undermine brand reputation as
well as business opportunities – causing businesses to fall behind competitors in the race to
unleash the vast business potential of AI.

Digital Business
AI enthusiasm outstrips deployment
Today, AI is clearly top of mind for many executives. Our study intended to demonstrate how this mindshare
translates into how and where companies are using the technology and the issues they face. Our findings
provide insight into the role that AI is playing and will play in companies, and how organizations will manage
and use the technology in the coming years.
Our study found that (no surprise) executives are enthusiastic about AI. They see it as having a growing
impact on their businesses, with 63% believing it is already extremely or very important to their company’s
success, and 84% saying that will be true in three years, including 48% who expect it to be extremely
important. Business leaders also believe it can be applied throughout their organizations. Respondents
reported on AI projects at their companies in a wide variety of functions, with customer service (30%),
manufacturing/production (26%), operations (20%) and research and development (18%) cited most often.
Moreover, executives have high expectations for AI. About three-quarters of respondents foresee
significant or major benefits from the technology, with one-third saying they expect major benefits. In
particular, 79% of respondents expect AI to introduce major or significant benefits in increased efficiency
and lower costs over the next three years (see Quick Take, page 5); 74% in increased revenues; and 73% in
improved ability to introduce new products and services, or enter new businesses. More than one-third
expect to see major benefits in each of these areas (see Figure 1 ).
Expected benefits of AI over next three years
37%
35%
37%
42%
39%
36%
79%
74%
73%
Increased efficiency/lower costs
Increased revenues
Ability to introducenew products/
services or enter new businesses
Major Benefit Significant Benefit
Response base: 975 executives in Europe and the U.S.
Source: Cognizant
Figure 1
Executives have high expectations for AI. About three-quarters
of respondents foresee significant or major benefits from the
technology, with one-third saying they expect major benefits.

Reimagining Digital
Content Services
We worked with a leading global K-12 publisher to accelerate its push toward digital
content creation and distribution, using a modern digital platform. The publisher’s existing
content operating model was distributed in silos globally, leading to long print cycles of
18 to 24 months.
With increased competition from digitally-savvy players, the company’s operations team
couldn’t keep pace with user demand for fresh content. Content reuse through multiple
delivery platforms — print, web and mobile — was seen as a way to satisfy customers while
reducing time to market.
Meanwhile, the company needed to manage a global vendor network of 170 content
producers, adding to the stress on operations and business competitiveness. With
content stored across 170 systems, it was difficult to derive optimal value from these assets
through content reuse, and the organization was unable to take advantage of a greater
collaborative opportunity to create content through enhanced workflows.
By applying our observations of industry trends and a deep understanding of the business,
we developed a solution premised on the following digital principles to produce and
manage content:
❙ Content is currency, and must be managed like treasury operations.
❙ Exemplary customer experience is a non-negotiable prerequisite.
Personalization is a must-have.
– Digitally-instrumented content operations can, and must, impact revenues.
5 / Making AI Responsible – And Effective Making AI Responsible – And Effective / 5
Quick Take
Digital Business
Accelerating fraud detection
While AI is still relatively new in business, some companies are already finding ways to
use it to increase efficiency and keep costs down. For example, in spite of a great deal of
automation in check processing operations, large banks still employ hundreds of people to
review computer images of scanned checks to spot signs of fraud. Not only is this process
time-consuming and inaccurate; it’s also ineffective, with banks losing millions of dollars
a year to counterfeiters.
Working with us, one global financial services organization took steps to use AI to augment
and streamline its fraud detection process. We developed a machine-learning solution
that essentially teaches itself to identify counterfeit checks. It analyzes scanned images
of handwritten checks, automatically comparing a variety of factors on the checks to a
growing database of checks that have been previously identified as fraudulent. The system
then flags potential counterfeits in near real-time while deposit transactions are in process,
helping to catch fraud early on.
When we tested our model on a historical portfolio of past transactions, it was able to
process up to 20 million checks per day, delivering a fast, accurate confidence score for
each check in less than 70 milliseconds. Fewer checks now require manual review, helping
to keep processing costs low. And the system is forecast to provide a $20 million reduction
in fraud losses annually, based on current models.1
Positive attitudes toward AI are particularly pronounced among respondents who said their companies
are growing much faster than the average company in their industries. Executives at these faster-growing
companies are more likely to view AI as important to their company’s success and more likely to expect
major benefits. If these leaders can unleash the potential of AI, their current lead could widen further (see
Quick Take, page 6).

Digital Business
Quick Take
The AI-growth connection
While executives in general are enthusiastic about AI, those from faster-growing
companies are especially interested in the technology – and that is only increasing the
urgency around AI implementations.
Among executives who said their company’s revenue growth is far above the industry
average, 45% said that AI is extremely important to the success of their companies today,
compared with only 14% of executives from slower-growing companies. Looking three
years ahead, that gap narrows, but it is still significant (63% vs. 43%) (see Figure 2, next
page).
Executives at faster-growing companies are also more likely to expect major benefits
from AI – and they expect to see them relatively soon, which suggests a more aggressive
approach to implementing these technologies. For example, these executives are more
likely than those at slower-growing companies to see major benefits in the next 12 months
in terms of increased revenue (36% vs. 11%), increased efficiency/lower costs (33% vs. 9%)
and their ability to introduce new products/services or enter new businesses (34% vs. 12%).
Executives from faster-growing companies also appear to be further along in terms of
exploring a wider range of AI technologies. Executives at these companies reported on
AI projects employing more key AI technologies: computer vision (64% vs. 47%), smart
robotics/autonomous vehicles (63% vs. 43%) and analysis of natural language (67% vs.
42%). And faster-growing companies are more likely than others to have ethics policies in
place for the initial design of AI applications (71% vs. 47%) and their behavior after they are
launched (70% vs. 43%).

AI embraced by fast-growers
Respondents were asked to rate the importance of AI to company success, today and in three years’
time.
45%
14%
41%
43%
86%
57%
Faster-growing companies
Slower-growing companies
Faster-growing companies
Slower-growing companies
63%
43%
29%
38%
92%
81%
IN THREE YEARS
Note: Faster-growing companies were deﬁned as those whose revenue growth is far above the industry average, as reported by respondents.
TODAY
Extremely important Very important
Source: Cognizant
Figure 2
Digital Business
These differences may be due in part to faster-growing companies’ ability to invest more in
new technology. Or it may be that they tend to be more innovative in terms of technology
usage – a key competitive quality today – and, therefore, more interested in the potential
of AI. In any case, these findings are especially important since investment in technology is
driving a trend in which larger companies are pulling further ahead of their smaller rivals.2
AI promises to take efficiency and productivity to new levels – and that makes it critical for
companies to develop and execute AI strategies sooner rather than later if they are not to
be left behind.

Digital Business
Current status of AI projects
21%
32%
48%
24%
Initial planning
Note: Percentages total to more than 100% since respondents reported on multiple projects.
Proof of concept
Pilot project
Full implementation
Source: Cognizant
Figure 3
Thereisadisconnectbetweenexecutives’enthusiasmfor
AIandtheiractualdeploymentofAIapplications.Inreality,
companies’real-worldexperiencewithAIisfairlylimited.
Attitudes toward AI vary somewhat by industry: Financial services, technology and retail executives are
most optimistic about both the importance and benefits of AI today, while insurance executives are less
so. But these differences are not especially large, and looking ahead, they nearly disappear, with about
eight out of 10 executives across industries saying AI will be extremely or very important to their company’s
success in three years. This broad, cross-industry appeal is fairly unusual for a new technology and is,
perhaps, a reflection of the wide variety of roles envisioned for the technology, as well as the tremendous
media attention it has received.
However, there is a disconnect between executives’ enthusiasm for AI and their actual deployment of AI
applications. In reality, companies’ real-world experience with AI is fairly limited. Only about two-thirds of
respondents were knowledgeable about an AI project at their company, and out of that group, only 24%
were knowledgeable about AI projects that were fully implemented. Most were aware of projects that
are still somewhat experimental: 21% reported on projects in the initial planning stage, 32% in the proof of
concept stage, and 48% in the pilot stage (see Figure 3). In short, companies appear to be moving fairly
slowly to implement actual AI projects.
This limited level of adoption is no doubt due in part to the relative newness of practical, business-ready AI
technology. But executives cited a range of challenges that they see with AI. Leading the list are the need
to secure executive talent, and attract and retain professionals with AI experience and skills, with 45% of
executives considering these factors to be extremely or very challenging when employing AI (see Quick
Take, next page). But not far behind are factors ranging from budget concerns (43%) to the interactions
between different AI applications (42%) and securing senior management commitment (40%) – a list that
may seem daunting to executives exploring the potential of AI.

Reimagining Digital
Content Services
We worked with a leading global K-12 publisher to accelerate its push toward digital
content creation and distribution, using a modern digital platform. The publisher’s existing
content operating model was distributed in silos globally, leading to long print cycles of
18 to 24 months.
With increased competition from digitally-savvy players, the company’s operations team
couldn’t keep pace with user demand for fresh content. Content reuse through multiple
delivery platforms — print, web and mobile — was seen as a way to satisfy customers while
reducing time to market.
Meanwhile, the company needed to manage a global vendor network of 170 content
producers, adding to the stress on operations and business competitiveness. With
content stored across 170 systems, it was difficult to derive optimal value from these assets
through content reuse, and the organization was unable to take advantage of a greater
collaborative opportunity to create content through enhanced workflows.
By applying our observations of industry trends and a deep understanding of the business,
we developed a solution premised on the following digital principles to produce and
manage content:
❙ Content is currency, and must be managed like treasury operations.
❙ Exemplary customer experience is a non-negotiable prerequisite.
Personalization is a must-have.
– Digitally-instrumented content operations can, and must, impact revenues.
Help still wanted
A number of observers today have expressed concern about the impact of AI on
employment, with AI’s potential to take on a variety of tasks traditionally performed by
people. But just how AI will affect the labor market is not that easy to predict. Indeed,
surveyed executives were divided on the question, with 37% saying AI will decrease the
number of people they employ, and 44% saying it will increase their employee ranks.
Retail and financial services industry executives were most likely to expect a boost to
employment (see Figure 4).
We believe that although AI will eliminate some jobs, others will emerge. We estimate that
over the next 10 to 15 years, nearly 21 million new jobs will be created by the advent of AI. In
addition, many employees will spend less time doing repetitive tasks and more time doing
tasks requiring judgment, empathy, creativity and human-specific skills.
AI will also create entirely new positions such as Chief Trust Officer or Ethical Sourcing
Manager. (Our “21 Jobs of the Future” report3
discusses these and other new jobs that are
likely to be created over the next 10 years.) That is, as work changes, new roles will emerge
9Digital Business
Quick Take
Expected impact of AI on company headcount over next three years
Decrease IncreaseNo change
Over Next Three Years
38%
48%
38%
38%
41%
29%
13%
22%
21%
22%
24%
15%
49%
30%
41%
40%
36%
56%
Financial services
Insurance
Healthcare
Manufacturing
Technology
Retail
Percentages may not round to 100% due to rounding
Response base: 596 executives in Europe and the U.S. who reported one or more AI projects
Source: Cognizant
Figure 4

10 / Making AI Responsible – And Effective10 / Making AI Responsible – And Effective
Digital Business
as others disappear. Surveyed executives seem to recognize this pending shift in work,
as most cited the need for additional skills and the need for employees to interact with AI
applications as important challenges.
To a great extent, tomorrow’s jobs will use AI to augment and complement human skills,
and use human-machine collaboration to bring together the best of each. Thus, we believe
that many new jobs will draw on human abilities for tasks such as coaching – helping others
do things well; caring – helping others with health and wellness; and connecting – helping to
connect people with people, people with machines, the physical with the virtual world, and
virtual work and commerce with ethics.4
Formulating an AI strategy
Companies will need to sharpen their focus and develop strategies for accelerating the move to AI. As
always, the world is not standing still: The technology is evolving quickly, and some early adopters are
already identifying AI opportunities and deploying AI technologies in the business.
The survey findings suggest that considerable uncertainty remains about how to proceed, and a lack of
strategic focus when it comes to AI. For example, while talent and skills top the list of challenges, executives
appear to give all challenges equal weight. Among the 13 issues listed in the survey, the difference
between the most-cited and least-cited was just seven percentage points, and roughly 40% of executives
considered each of the 13 issues to be extremely or very challenging. This “all of the above” finding suggests
that companies have not developed clear strategies for pursuing AI. So too does the nature of some of
their challenges, such as securing adequate budget (43%), senior management commitment (40%) and
buy-in by the business (38%).
This lack of strategy is further underscored by the fact that roughly the same percentages of executives
cited specific AI technologies as being used in their projects: virtual agents (55%), computer vision (51%),
advice engines/machine learning (49%), smart robotics/autonomous vehicles (48%) and analysis of
natural language (48%). As with the AI challenges, these responses once again essentially add up to “all of
the above.” This indicates that companies are not yet at a point of understanding where and how specific
technologies can benefit the business.
To create rigorous AI strategies, companies need to look beyond technological capabilities. There is no
recipe that companies can use to embed AI in the fabric of the company – each business challenge will
require different tools, techniques and approaches. Rather than a sequential process, leveraging AI will
demand extensive experimentation and the ability to apply learnings to the next stage of deployment.
Companies need to factor that reality into their plans.
AI strategy should start with an emphasis on the creation of business value and ethical/responsible
behavior. That may sound like common sense, but many AI initiatives today are focused more on the
technology’s capabilities and algorithms than on its impact and benefit for the business. That is, they tend
to ask questions such as, “What can IBM’s Watson or Amazon’s machine learning do?” rather than, “What
business problems or opportunities do we have that could be addressed with AI?”

Digital Business
Digital Business
Companies often think of AI as a tool for reducing costs, and that is certainly a key benefit – one that tops
the list for executives in the survey. But there is a range of other potential business benefits to consider.
If done right, AI can help improve product and service quality; reduce cycle time; create new and better
employee experiences; and enhance safety, among other things.
Employing AI effectively requires a clear focus on applying intelligent technologies to solve tough
operational challenges and deliver a lift to the business. For example, we worked with a utility to deploy
voice-activated, AI-driven chatbots that assist executives, account managers and field-service technicians
in conducting research into services and solutions using voice commands or typed queries, which helped
the utility streamline customer interactions and enhance user experience. In another project, we helped
an insurer use machine learning, along with geospatial analysis, to better understand the complex flood-
insurance market – an effort that identified an overall financial opportunity of $3.3 billion and opened the
door to a new line of business for the company.
Each company’s situation is different, of course. But there are some general guidelines that companies can
follow in the search for AI-enabled business value:
❙ Look for opportunities to leverage data. Companies are increasingly awash in data about everything
from customer preferences and competitors to the details of Internet of Things (IoT)-equipped
production facilities. This flood of data can quickly overwhelm humans trying to understand and
interpret it. The irony: As more data is generated in the modern digital era, separating signal from noise
becomes more difficult. AI, with its ability to apply human-like assessments and decisions to that data,
rapidly and efficiently, provides a possible solution to that dilemma (see Quick Take, page 13).
❙ Cast a wide net. AI has the potential to touch many parts of the company, and it’s important that
underlying algorithms “understand” the larger context in which they operate. If they don’t, they are
likely to make recommendations for the right product but at the wrong time, for example, or struggle
to make sound decisions involving cost and speed tradeoffs. With that in mind, organizations should
establish cross-functional teams to identify AI opportunities, and arm them with a structured, systematic
approach to uncovering potential improvements to processes and stakeholder relationships.
❙ Solicit input and insight from external parties. Partnering can be an important factor in AI strategy. In
our survey, executives most often cited access to AI skills as a key challenge, and working with vendors
can allow companies to quickly gain access to the required skills. Many companies appear to recognize
this: 62% of executives said they rely on a mix of vendor and in-house resources for AI, while 13% said
they always or usually secure AI applications from vendors. Companies should consider expanding
these relationships into robust ecosystems that can work collaboratively to keep up with rapidly evolving
technologies and use cases, and accelerate the journey to AI.

Digital Business
❙ Encourage experimentation – and discipline. While AI is rapidly moving into real-world business
applications, it is still an emerging technology. To a great extent, companies will need to chart their own
course forward – there are no ready-made, turnkey solutions, and there is still much to learn. That makes
managed experimentation key. Companies should encourage a tolerance for risk-
taking and innovation with AI, but balance that with rigorous testing and measurement of ROI and
tangible business value. They should move on from failures quickly, and at the same time, be prepared to
quickly scale up successful experiments and pilots into production.5
Perhaps most important, an effective strategy requires a human-centric approach to AI. The technology is
most powerful when it collaborates with people and augments human activities and
decisions – and a human-oriented approach is key to ensuring that AI fits into that real-world business
context (see Quick Take, page 15). Thus, companies should consistently ask how AI can enable and support
employees and customers, how it can best interact with people, how it can be designed to understand the
behaviors and expectations of humans, and how it can have enough “emotional intelligence” to adjust its
decisions to the people around it.
To help ensure a human-centric approach, companies can complement the input of business people
with the insights of behavioral scientists (sociologists, anthropologists, etc.) to shape and train AI to be an
effective, and valued, part of the organization.
AI technology is most powerful when it
collaborates with people and augments human
activities and decisions.

Working smarter to reduce
compliance risk
AI brings a fundamental quality to the table – the ability to help companies leverage
today’s growing volumes of data, quickly and intelligently. That proved to be critical
to an international professional services organization’s efforts to reduce financial and
compliance risk. The organization frequently has to conduct due diligence on the
numerous vendor, partner and counterparty companies that are part of its business
ecosystem. This typically involves going through more than 40,000 global sources,
including media reports, corporate records, financial transactions and legal cases. This
information had to be painstakingly reviewed for each entity before a report could be
finalized – a laborious process that could take weeks.
The organization worked with us to apply advanced forms of AI to tackle the problem.
It had adopted a third-party research tool to compile search results, but that system
lacked “smart” search capabilities, and researchers were still spending a large amount of
time reviewing irrelevant results. To improve the functionality of the system, we quickly
developed a proof of concept for a machine-learning platform based on IBM’s Watson
Explorer 11 and Watson Content Analytics, along with advanced APIs to link the AI platform
to the firm’s existing systems.
With the solution, a contextual search model allows the organization to detect and
deliver improved reports on news relating to risk and fraud. A machine-learning model
understands linguistic nuances, meanings and relationships specific to an industry. This
model extracts relevant information from enormous, diverse data sets, producing research
reports with contextually relevant results. The solution can rank results by the likelihood of
risk based on context, without researchers having to conduct multiple searches.
13Digital Business
Quick Take

Digital Business
The solution automates more of the search process, integrates research workflow and
reduces the time spent manually reviewing material by more than half – and 14% of reports
can now be completed in one hour. As a result, the organization can generate up to
30% more due diligence reports a year, helping to reduce compliance and financial risks
significantly.
6
Digital Business

Bringing human-centric AI
to healthcare
When approached from a human-centric perspective, AI can benefit individuals,
businesses and society as a whole. For example, opioid dependency is devastating for
patients and their families, and it can limit positive outcomes when addicted patients are
being treated for other illnesses. It’s also expensive: U.S. healthcare organizations spend
more than $500 billion annually to care for patients suffering from opioid addiction, and
treating addicted patients diverts much-needed resources from other patients.7
We worked with a leading healthcare services provider to develop an AI-driven solution
that detects drug-seeking behavior and identifies patients who are at risk of becoming
addicted. The solution uses an advanced machine-learning algorithm to mine information
about patient behaviors and symptoms, and examines physicians’ notes from patient
interactions, captured using text analytics.
The system combines phase-based extraction, rule-filtering and advanced text clustering
to mine highly variable data and ultimately identify patients who could become drug-
seekers – before they turn into addicts. Pop-up alerts in the organization’s electronic
medical records system prompt physicians to take corrective actions at the point of care,
interceding with patients in real time. The solution also learns continuously from its own
results to verify the accuracy of its models and improve searches.
The tangible health and financial benefits from this solution include saving the healthcare
organization as much as $60 million through the identification of some 85,000 at-risk
patients. The intangible, but nevertheless invaluable, benefits include improved quality of life
for thousands of patients, their families and society.
15Digital Business
Quick Take

Digital Business
Overcoming strategic challenges
Even the most well-developed AI strategies won’t succeed unless fundamental challenges are considered.
One of these challenges is the ability to build a profitable and scalable business from AI. Many established
businesses start with foundational AI centers of excellence, but these teams quickly confront questions
that extend beyond the improvement of customer or user experience, or the use of AI to automate a
process, such as:
❙ Can they trade in AI services in the future?
❙ Can they hive off a profitable business based on AI services on top of their existing product or service?
A large home appliance company recently partnered with a number of start-ups to understand how people
cook, with the objective of influencing cooking and improving the dining experience. The company wanted
to create a platform that will help it accomplish this goal using AI microservices embedded in its devices.
Eventually, the company foresees an operating model where it monetizes this microservice by engaging
retail partners across areas such as food, delivery and packaging. Other traditional companies are also
partnering or buying AI start-ups in anticipation of realizing a “profitable and scalable” business from AI.
Another challenge is scalability of AI microservices. For example, each connected car or connected home
appliance will have several different AI microservices and deep-learning models running on its units/
devices to deliver a variety of services. Accommodating the millions of devices in use that want to make use
of these microservices, globally, is non-trivial. For digital-native companies, this is slightly more manageable
because they own more of their value chains and often have built their business around intelligent
algorithms. However, the challenge for traditional companies like automobile manufacturers and home
appliance makers is that their AI services scope will be incomplete unless they tap into their traditional
value chain, including manufacturing, logistics, service partners, and warranty and fulfillment partners, to
effect a wider AI impact.

Digital Business
AI governance: building transparency, trust and personalization
For companies to successfully unleash the potential of AI, people will need to see it as a reliable,
dependable colleague. It will interact with customers, employees and partners, running operations
and making important decisions. If an AI application is not well designed and managed, it may end up
“misbehaving.” The ramifications could be significant, ranging from damaged customer relationships, to
missteps in factories that affect quality, to discriminatory decisions that elicit regulatory scrutiny.
Companies will need to put the right AI governance structures in place to guide its design and use. They’ll
need to develop policies for validating and operating the AI system, and work continuously to monitor
and correct AI behavior. To that end, companies should consider establishing an AI council to monitor and
oversee AI, and focus on ensuring three key factors: transparency, trust and personalization.9
Transparency: beyond the black box
Transparency boils down to allowing people to understand how an AI application makes decisions.
Currently, AI systems tend to be “black boxes” whose operations are not well understood by humans,
and that makes some people uncomfortable. We want to know why our colleagues and friends reach the
conclusions they do, and we would like to do the same with intelligent machines. Companies will need to
help stakeholders understand what AI applications know about people and why they behave the way they
do. For example:
❙ Banking customers will want to know the criteria systems use for approving loans.
❙ Plant operators will want to know why AI decided to shut down a production line for safety reasons.
❙ On a more mundane level, customers will want to know why they are being offered, say, hiking boots
rather than running shoes.
Transparency helps build trust, but trust is based on other factors, too. Can people feel confident that
the system is working as it was designed to work, and that it understands their needs? Consistency, too, is
important – people learn to trust over time, through experience.
Companies will need to put the right AI governance
structures in place to guide its design and use.
They’ll need to develop policies for validating and
operating the AI system, and work continuously to
monitor and correct AI behavior.

18 / Making AI Responsible – And Effective18 / Making AI Responsible – And Effective
Trust: it starts with the data
High-quality data is key to earning trust. Data needs to be not only accurate and free of “noise,” but also
free of bias. AI is not programmed in the traditional sense; rather, it learns from examples – and biased data
is essentially a bad example. That was clearly illustrated by Microsoft’s “Tay” chatbot, an AI application that
was intended to learn how to behave well by interacting with Twitter users but ended up producing racist
comments after a few hours of listening to online trolls.10
On another level, researchers have noted that lending and credit-scoring AI applications that learn from
historical data can quickly pick up the gender and race biases built into that data. Companies will need
to find ways to systematically mitigate biases and ensure that AI is learning from the right inputs – or
risk implementing the kind of error-laden systems that a World Economic Forum report called “artificial
stupidity.”11
(Such issues can also raise ethical questions; see next section, next page).
In using data with AI, companies will need to keep compliance in mind, as regulations around data privacy
continue to evolve and expand. For example, the European Union’s General Data Protection Regulation
(GDPR) is a recently enacted set of rigorous data privacy rules designed to safeguard the privacy rights of
data subjects – that is, customers and others whose data is used by a company. These rules are specific to
AI, but will have important ramifications for the field. For example, GDPR requires companies to:
❙ Institutionalize the practice of asking data subjects for consent to use their data.
❙ Make the logic behind automation clear to data subjects who ask about it.
❙ Establish data protection checks and balance so that AI is not drawing on data that goes beyond its
stated purpose.
While the GDPR is a European Union regulation, it applies to companies headquartered anywhere that do
business in Europe, including U.S. companies.
Personalization: making it relevant and useful
Finally, personalization will be important in driving the acceptance and success of AI, because it is key to
performing tasks and making decisions that are relevant and useful for the people they impact. Employees,
for example, typically tailor their interactions with one another based on the individuals and situation
involved. AI applications will need to do the same, whether it is suggesting an action to a banker or
recommending a product to a customer.
This shaping of interactions not only increases the effectiveness of AI; it can also enhance the trust that
people feel for the technology.
Data needs to be not only accurate and free of “noise,”
but also free of bias. AI is not programmed in the
traditional sense; rather, it learns from examples – and
biased data is essentially a bad example.

Digital Business
Do the right thing: responsible, ethical AI
In the end, the goal of good governance is “responsible AI.” If AI is not responsible, it won’t be embraced
by employees or customers, not to mention shareholders who are put off by potential legal and
reputational risk.
There is another, more fundamental thread at the heart of good AI governance and responsible AI: ethics.
Increasingly, the development of AI ethics is recognized as critical to the success of AI, and it needs to be
interwoven into everything companies do with the technology.
That means not only building ethical AI systems, but also running those systems ethically, in ways that
are aligned with corporate and societal values. By its nature, AI is always adapting to and building on new
information. Ethics will become particularly important as AI becomes more ubiquitous, and as AI systems
increasingly learn from one another, not just from the inputs that humans provide – that is, when machine-
learning AI applications “teach” other AI applications. An ethical foundation is needed to ensure that
AI continues to operate for the benefit of humans over time. In a very real sense, ethics is at the heart of
human-centric AI.
That reality is driving a growing number of programs that are focused on AI ethics. For example, the MIT
Media Lab and Harvard University’s Berkman-Klein Center for Internet and Society have launched an
Ethics and Governance of AI Initiative to look at AI and the “social values of fairness, human autonomy and
justice.”14
The University of Oxford is collaborating with other UK universities to find ways to make AI more
transparent and accountable.15
Facebook, Google, Microsoft and other tech companies have created The
Partnership on AI to explore AI best practices and conduct “open research and dialogue on the ethical,
social, economic and legal implications of AI.”14
And New York City recently established a task force to look
at fairness and accountability in AI decisions made in various city departments.15
Companies will need to find ways to train AI to behave ethically. This is not as arcane a process as one might
think. The concept is familiar to parents who provide feedback and guidance to raise their children to be
good members of society, and there are well-understood tools and frameworks from the world of human
sciences that can be used to instill ethics into the design and operation of AI.
While AI ethics is becoming an increasingly high-profile concern, its importance is not always
widely recognized in business. Less than half of the executives (45%) surveyed said that ethical
Ethics will become particularly important as AI becomes
more ubiquitous, and as AI systems increasingly learn from
one another, not just from the inputs that humans provide –
that is, when AI applications “teach” other AI applications.

Digital Business
considerations play a critical or significant role when their company develops and employs AI. In contrast, a
majority (55%) said that ethical considerations play some, a minor, or little or no role.
Only about half of respondents reported that their companies have policies and procedures in place to
address potential unethical behavior in the initial design of AI, although 35% plan to put these in place.
Similar percentages of executives said the same about their approach to dealing with unethical behavior
after AI applications are launched. In terms of how they uncover unethical behavior, executives most often
cited testing by employees (59%), followed by customer feedback (54%), technology tools provided by
external vendors (49%) and custom-built technology tools developed in-house (47%) (see Figure 5).
Approximately two-thirds of executives believe their companies’ policies for ensuring ethical AI behavior
are extremely or very effective. This somewhat optimistic view may be based in part on the relatively limited
rollout of AI at many companies. As AI finds its way into more parts of the organization, and as it continues
to learn and evolve in real-world business situations, maintaining an ethical foundation will become more
complicated as well as more important. We believe that in time, many companies that are satisfied with
their current AI ethics policies and procedures will realize they need to redouble their efforts in order to
keep up with a growing range of ethical questions, such as:
❙ Who bears legal responsibility for a machine’s harmful conduct?
❙ Should people be made aware when they are dealing with a very human-like AI system?
❙ What oversight is needed to keep AI from being used for criminal purposes?
❙ How can we manage the economic and labor disruptions that AI might bring?
❙ Should limits be placed on the intelligence and autonomy of AI applications?
To answer these questions, companies will need to maintain a relentless focus on ethics over time.
Tools used to identify and address potential unethical
behavior in AI applications
59%
54%
49%
47%
Testing by employees
Customer feedback
Technology tools provided
by external vendors
Custom-built technology
tools developed in-house
Multiple responses permitted
Source: Cognizant
Figure 5

Digital Business
Looking forward
Unlike many established information technologies, it is often easy to get started with AI. New AI systems
can be developed fairly quickly – an experimental version can typically be up and running in a month, and
pilots can be rolled out in a month or two after that. The trick is to move rapidly from that starting point to
grow and sustain full-scale, business-ready AI applications.
To help move projects forward, companies should start with an assessment of how prepared the organization
is for AI, and what it needs to be put in place. What platforms will be used to run the scaled-up application?
What skills and knowledge will be needed, and will these come from in-house personnel or be sourced from a
partner? What data will be needed from traditional transaction systems, such as sales, marketing and finance;
from newer digital systems, such as point-of-sale and social media applications; and from external, context-
setting sources, such as governments, weather services, online review sites and so forth?
In terms of launching projects, companies may want to focus their initial AI efforts on targeted back-
office applications, an approach that can provide an opportunity to gain experience with the technology
without risking high-profile problems with customer-facing systems. They should also look for quick-win
opportunities that allow them to chalk up successes and financial benefits to build AI momentum. And
they need to understand and address AI’s impact on employees – how their roles are shifting and what new
skills they will need – and then employ change-management techniques to help people succeed with new
AI-based tools.
As they nurture AI and move it from experimentation to the business core, companies should also keep an
eye on the big picture. That will mean formulating plans for governance, and always keeping the need for
responsible AI top of mind. Through it all, they should look at their efforts through a human-centric lens –
because that will be key to designing and reaping business value from AI in the long run.
New AI systems can be developed
fairly quickly. The trick is to move
quickly from that starting point
to grow and sustain full-scale,
business-ready AI applications.

Respondent industry
Financial services
24%
Technology
21%
Retail
14%
Healthcare life sciences
12%
Insurance
5%
Media entertainment
3%
Manufacturing
21%
Digital Business
Methodology
We conducted a survey, through a combination of online and telephone interviews, with 975 respondents
from companies in the U.S. and Europe in June and July 2018.
The companies represented a variety of industries and sizes, as measured by their annual revenues.
Fifty percent of the respondents were from companies headquartered in the U.S., 48% from companies
headquartered in Europe, and 2% from companies in other locations. The respondents from European
companies were located in the following countries: UK (11%), Germany (11%), France (10%), Netherlands
(5%), Sweden (4%), Belgium (4%) and Norway (3%).
More than half the respondents had C-suite titles, with an additional 16% being an executive vice-
president, senior vice-president or vice-president. Respondents work in the following functional areas:
information technology (39%), management (19%), finance (13%), operations/manufacturing (9%),
research development (4%), sales marketing (4%), human resources (4%), customer service (3%) and
other functions (5%).

Annual revenues
$10 million – $100 million
7%
$100 million – $500 million
33%
$500 million – $1 billion
28%
$1 billion or more
32%
Respondent title
C-suite
54%
EVP/SVP/VP
16%
Senior director/director
25%
Senior manager/manager
5%
Digital Business
Digital Business

Endnotes
1 For more information on this and other case studies related to employing AI to deliver business value, see “Accelerate
Business Growth and Outcomes with AI,” Cognizant Technology Solutions, Sept. 7, 2017, https://www.cognizant.com/
perspectives/accelerate-business-growth-and-outcomes-with-ai; or visit the “Featured Work” section of Cognizant’s AI
web pages: https://www.cognizant.com/artificial-intelligence.
2 Christopher Mims, “Why Do the Biggest Companies Keep Getting Bigger? It’s How They Spend on Tech,” The Wall Street
Journal, July 26, 2018, https://www.wsj.com/articles/why-do-the-biggest-companies-keep-getting-bigger-its-how-they-
spend-on-tech-1532610001.
3 “21 Jobs of the Future,” Cognizant Technology Solutions, November 2017, https://www.cognizant.com/whitepapers/21-
jobs-of-the-future-a-guide-to-getting-and-staying-employed-over-the-next-10-years-codex3049.pdf
4 Ibid.
5 For a discussion on developing an AI strategy, see “The Road to AI,” Cognizant Technology Solutions, August 2018. https://
www.cognizant.com/whitepapers/the-road-to-ai-codex3614.pdf.
perspectives/accelerate-business-growth-and-outcomes-with-ai, or visit the “Featured Work” section of Cognizant’s AI
7 “The Underestimated Cost of the Opioid Crisis,” The Council of Economic Advisers, November 2017, https://www.
whitehouse.gov/sites/whitehouse.gov/files/images/The%20Underestimated%20Cost%20of%20the%20Opioid%20
Crisis.pdf.
perspectives/accelerate-business-growth-and-outcomes-with-ai; or visit the “Featured Work” section of Cognizant’s AI
9 For a discussion of AI governance, see “AI: Ready for Business,” Cognizant Technology Solutions, July 2018, https://www.
cognizant.com/whitepapers/ai-ready-for-business-codex3752.pdf .
10 “Bias In, Bias Out: How AI Can Become Racist,” Design News, Sept. 29, 2017. https://www.designnews.com/content/bias-
bias-out-how-ai-can-become-racist/176888957257555.
11 “Top 9 Ethical Issues in Artificial Intelligence” World Economic Forum, Oct. 21, 2016, https://www.weforum.org/
agenda/2016/10/top-10-ethical-issues-in-artificial-intelligence/.
12 The Ethics and Governance of Artificial Intelligence Initiative website: https://aiethicsinitiative.org/.
13 “Oxford to Be Key Player in New AI Accountability Project,” University of Oxford, May 9, 2018, http://www.ox.ac.uk/
news/2018-05-09-oxford-be-key-player-new-ai-accountability-project.
14 Partnership on AI website: https://www.partnershiponai.org/tenets/.
15 “Mayor de Blasio Announces First-In-Nation Task Force To Examine Automated Decision Systems Used By The City,” New
York City press release, May 16, 2018. https://www1.nyc.gov/office-of-the-mayor/news/251-18/mayor-de-blasio-first-in-
nation-task-force-examine-automated-decision-systems-used-by.
Digital Business

Digital Business

About the author
Bret Greenstein
Global Senior Vice President Head of Digital Business
AI Practice, Cognizant
Bret Greenstein is Global Senior Vice President and Head
of Cognizant’s Digital Business AI Practice, focusing on
technology and business strategy, go-to-market and
innovation, helping clients realize their potential through digital
transformation. Prior to Cognizant, Bret led IBM Watson’s
Internet of Things offerings, establishing new IoT products
and services for the industrial Internet of Things (IIoT). He built
his career in technology and business leadership across a range of roles throughout IBM, in
software, services, consulting, strategy and marketing, and served as IBM’s CIO for the Asia-
Pacific region. Bret has worked globally in these roles, including living in China for five years,
working with clients and transforming IBM’s IT environment. Bret holds patents in the area of
collaboration systems. He holds a bachelor’s degree in electrical engineering and a master’s
degree in manufacturing systems engineering from Rensselaer Polytechnic Institute. He can be
reached at Bret.Greenstein@cognizant.com | www.linkedin.com/in/bretgreenstein/.
Acknowledgments
The authors would like to thank Matt Smith, an AVP in Cognizant’s Conversational AI Practice, and
Abhishek Rishi, a Director in Cognizant Consulting’s AI Analytics Practice, for their invaluable
contributions to this study.
Digital Business
Digital Business

Digital Business

© Copyright 2020, Cognizant. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, transmitted in any form or by any means,electronic, mechanical,
photocopying, recording, or otherwise, without the express written permission from Cognizant. The information contained herein is subject to change without notice. All other trademarks
mentioned herein are the property of their respective owners.
Codex 3974.04
World Headquarters
500 Frank W. Burr Blvd.
Teaneck, NJ 07666 USA
Phone: +1 201 801 0233
Fax: +1 201 801 0243
Toll Free: +1 888 937 3277
European Headquarters
1 Kingdom Street
Paddington Central
London W2 6BD England
Phone: +44 (0) 20 7297 7600
Fax: +44 (0) 20 7121 0102
India Operations Headquarters
#5/535 Old Mahabalipuram Road
Okkiyam Pettai, Thoraipakkam
Chennai, 600 096 India
Phone: +91 (0) 44 4209 6000
Fax: +91 (0) 44 4209 6060
APAC Headquarters
1 Changi Business Park Crescent,
Plaza 8@CBP # 07-04/05/06,
Tower A, Singapore 486025
Phone: + 65 6812 4051
Fax: + 65 6324 4051
About Cognizant Digital Business
Cognizant Digital Business helps our clients imagine and build the Digital Economy. We do this by bringing together human insight, digital strategy,
industry knowledge, design, and new technologies to create new experiences and launch new business models. For more information, please visit
www.cognizant.com/digital or join the conversation on LinkedIn.
About Cognizant AI Analytics Practice
As part of Cognizant Digital Business, Cognizant’s AI Analytics Practice provides advanced data collection and management expertise, as well as
artificial intelligence and analytics capabilities that help clients create highly-personalized digital experiences, products and services at every touchpoint
of the customer journey. We apply conversational AI and decision support solutions built on machine learning, deep learning and advanced analytics
techniques to help our clients optimize their business/IT strategy, identify new growth areas and outperform the competition. Our offerings include AI to
Insight, Customer Intelligence, Intelligent Automation, Product Intelligence, and Risk Fraud Detection. To learn more, visit us at www.cognizant.com/
cognizant-digital-business/applied-ai-analytics.
About Cognizant
Cognizant (Nasdaq-100: CTSH) is one of the world’s leading professional services companies, transforming clients’ business, operating and technology
models for the digital era. Our unique industry-based, consultative approach helps clients envision, build and run more innovative and efficient business-
es. Headquartered in the U.S., Cognizant is ranked 193 on the Fortune 500 and is consistently listed among the most admired companies in the world.
Learn how Cognizant helps clients lead with digital at www.cognizant.com or follow us @Cognizant.

Making AI Responsible and Effective

More Related Content

What's hot

Similar to Making AI Responsible and Effective

More from Cognizant

Making AI Responsible and Effective