Explore real-world examples of AI implementation success

In 'All-In on AI,' authors Davenport and Mittal explore AI implementation examples from organizations that already made the AI leap with success. Read this book excerpt to learn more.

Although many organizations are interested in implementing AI, leaders are often unsure where to start or how much to invest in the technology. Should they wade cautiously into the water or take a full leap into the deep end?

The book cover for 'All-In on AI: How Smart Companies Win Big With Artificial Intelligence.'Click the book cover
to learn more.

In their book All-In on AI: How Smart Companies Win Big With Artificial Intelligence, published by Harvard Business Review Press, authors Thomas H. Davenport and Nitin Mittal offer a glimpse into a small group of companies that have taken the gamble of fully integrating AI into their processes and business strategies -- and gained a competitive advantage as a result. Some of the most important aspects of AI implementation, Davenport and Mittal explain in the book, are human leadership, behavior and change.

The authors argue that organizations must make large investments in AI to see big results; the technology is nothing if not everybody is on board. "To achieve substantial value from AI, a company should fundamentally rethink the way humans and machines interact within working environments," they write in the first chapter of the book.

In the following excerpt from "Chapter 2: The Human Side" of All-In on AI, Davenport and Mittal dive into the human side of AI integration, exploring the benefits of upskilling or retraining employees to better equip them for impending job changes. Through real-world examples, they explain how organizations are approaching AI education and data science training.

Perhaps the most logistically challenging among the human side of AI issues is educating employees about its capabilities and likely future impact on their jobs. It's difficult for a variety of reasons: there are many employees within a large organization; it's hard to predict what changes in jobs will take place because of AI over the next several years; and finally, different employees have different objectives and interests relative to their jobs, so "one size fits all" educational initiatives are unlikely to be successful.

Some companies -- generally not those that are all-in on AI -- have taken these challenges as a reason to limit educating their employees about AI. One large defense contractor's HR leadership, for example, justified their approach with three arguments:

  1. The company has many other competing priorities in the near term. Is it worth investing in something that is so long term and uncertain in its impact?
  2. Job changes and automation are moving a lot more slowly than the experts predicted. We'll be able to adjust as the changes come. When jobs do change, most of the time it's task augmentation or new skills rather than layoffs. That kind of change is less difficult to accomplish and easier to plan for.
  3. There's so much uncertainty around the prognostication that we're likely to be wrong. Then the company will need to do real-time adjusting anyway.

Although these arguments are reasonable, we take a different view. We believe it is possible to predict some changes in jobs from AI, or at least to better equip employees to prepare for more generic job changes. And while it's true that augmentation is more likely than large-scale automation, augmentation will likely lead to job changes for which workers need to be prepared. Our 2018 survey of AI adopters found that 82 percent expected moderate or substantial job changes for their employees in three years. Despite competing priorities, we believe the time is now to educate employees about AI and its impacts. It's likely to take a while, so there is little time to waste. These happen to be the same ideas that some AI-focused companies are using to justify their present actions.

Of course, some organizations that want to retrain or upskill workers aren't certain what specific skills will be required for jobs of the future, but they are confident that those skills will be digitally oriented. Amazon, for instance, has committed to spend $700 million on retraining to ensure that its employees have the skills they will need to thrive in an increasingly digital job market -- within or outside Amazon. The company's primary focus is the third of its workers in distribution centers, its transportation network, and nontechnical roles at headquarters. It provides retraining for workers in distribution centers (which are more vulnerable to automation) for jobs as IT support technicians, and for nontechnical corporate workers in software engineering skills.

Similarly, leaders at DBS Bank in Singapore provided employees with seven digital skills, including digital communications, digital business models, digital technologies, and data-driven thinking. The program is called DigiFY, and it is aimed at upskilling many of the bank's employees. Deloitte has focused on making its professionals tech savvy -- assuming that in an AI-oriented business environment, virtually every employee will need to understand how technology works and fits with their jobs. All three companies believe that whatever changes happen to future jobs, employees -- and their employers -- will be better off if they are more skilled at digital technologies.

Sometimes these new skills lead to a new set of roles. DBS also created a group of "translators" -- people who are quantitatively oriented, but not data scientists, and who can mediate between business stakeholders and AI developers. This role is an important one that has been fairly widely discussed, but not widely implemented. DBS has even decided to staff AI projects with one translator for every two data scientists. Sameer Gupta, the bank's Chief Analytics Officer, said that when the two roles collaborate, data scientists can be more experimental with their modeling, while the translator can ensure that the actual business problem is being specifically addressed.

A variation on this strategy is to educate employees in data science skills. This approach often involves working with the providers of online courses in that area. Shell, for example, began a relationship with Udacity in 2019, when the energy giant realized it had nowhere near the number of data scientists needed to complete all the AI-related projects it planned. It created a pilot program for people with IT backgrounds and then embarked on a larger initiative aimed at petroleum engineers, chemists, data scientists, and geophysicists, among others. Completing the AI nanodegree typically takes four to six months to finish, working ten to fifteen hours per week. As of this writing, more than five hundred employees had completed or were currently enrolled in the nanodegree program, and an additional one thousand staff had completed data literacy and digital literacy courses.

Similarly, Airbus has partnered with Udacity to train more than a thousand employees in data science and analytics. The company asks both employees and their managers to devote half a day a week to the training. Managers work with employees to identify a pilot project in data science the employees can work on, and the managers monitor their progress. Airbus believes the training initiative has multiple benefits. It has not only increased the number of people who can work with AI but has also led to a community of people who are interested in data science and AI with whom the central data science group can collaborate. The training program is also a means to deploy AI best practices around the company, and the projects are a way to familiarize managers and their businesses with AI.

Reprinted by permission of Harvard Business Review Press. Excerpted from All-In on AI: How Smart Companies Win Big With Artificial Intelligence by Thomas H. Davenport and Nitin Mittal. Copyright 2023 Deloitte Development LLC. All rights reserved.

About the authors
Thomas H. Davenport is the President's Distinguished Professor of information technology and management at Babson College, a visiting professor at Oxford's Saïd Business School, a research fellow at the MIT Initiative on the Digital Economy and a senior adviser to Deloitte's analytics practice. His bestselling books include
Competing on Analytics and Big Data at Work.

Nitin Mittal is a principal with Deloitte Consulting LLP. He currently serves as the U.S. AI strategic growth offering consulting leader and the global strategy, analytics, and mergers and acquisitions practice leader.

Next Steps

Evaluate model options for enterprise AI use cases

Dig Deeper on AI business strategies

Business Analytics
Data Management