AI is undoubtedly booming in business—at least in certain segments of it.  In my work, I’ve helped design and analyze surveys suggesting that 25% to 30% of large U.S. companies are pursuing AI, many quite aggressively. Some have hundreds or even thousands of projects underway. The firms using AI most aggressively are large businesses with the most data—online platforms, financial services, telecommunications, and retail. Small- to medium-sized enterprises, business-to-business firms, and those in basic manufacturing industries are less likely to use AI. They typically lack not only the data to succeed with AI, but the expertise and awareness to pursue it effectively. Firms outside the United States are also pursuing AI at a slower pace, although there are aggressive adopters in China, the U.K., Canada, and Singapore.

A variety of different AI technologies are in use.  You need to be aware of which ones do what. As Emma Martinho-Truswell explains in her article, machine learning is perhaps the most important component of AI, but it has multiple variations—ordinary statistical machine learning, neural networks, deep learning neural networks, and so on. Versions of AI also use semantic approaches to understanding language and logic-based rule engines for making simple decisions. Each technology performs a particular set of tasks; deep learning, for example, excels at recognizing images and speech.

AI is being applied for various business purposes.  The most common uses enable us to make better decisions, improve operational processes, and enhance products and services. The first two are an extension of business analytics and typically employ machine learning; product-oriented objectives are common in high-tech firms, automobiles, and advanced manufacturing.

Many large companies are creating infrastructures and processes to manage AI.  More than a third of large U.S. firms report in multiple surveys that they have a strategy in place for AI, have created a center of excellence to facilitate its use, and have identified its champions in the management team. As Vikram Mahidhar and I suggest in our article, late adopters may have difficulty catching up.

Companies are finding success by focusing their AI efforts in certain areas of the organization.  Given the combination of short-term incremental value and long-term opportunity, many companies are tempering expectations about AI while still providing motivation to move forward aggressively with the technology. This is perhaps best accomplished by undertaking several projects focused in a particular area rather than spreading AI projects throughout the organization. Transforming customer service, for example, might include projects involving chatbots, intelligent agents, recommendation engines, and so forth.

AI hasn’t transformed business—yet.  While surveys suggest high expectations for transformation and high percentages of respondents say they have achieved economic returns, there are few examples of sweeping business reinvention thus far, for several reasons:

• It’s still early in the life cycle of AI activity.
• Not every company has data that’s suited for AI, as Ajay Agrawal, Joshua Gans, and Avi Goldfarb explain in their article (although H. James Wilson, Paul Daugherty, and my son Chase Davenport suggest in their article that data requirements for effective AI may lessen in the future).
• Companies are undertaking pilots with AI rather than production deployments, as Andrew Ng recommends in his piece.
• AI tends to be a narrow technology that supports particular tasks, not entire jobs or processes.
• Highly ambitious moon-shot projects, such as treating cancer, enabling autonomous vehicles, and powering drone deliveries, have been unsuccessful or slow to arrive.

Even at data powerhouses like Amazon, most AI activity has involved projects that “quietly but meaningfully improve core operations,” according to CEO Jeff Bezos in his 2017 letter to shareholders. It’s an evolutionary set of improvements that will eventually add up to revolution.

AI’s overall impact on employment isn’t certain, but jobs will clearly change.  Some observers have predicted dire levels of AI-driven unemployment. Thus far—as Wilson and Daugherty discuss in their article on “collaborative intelligence”—augmentation of human work by smart machines has been far more common than large-scale automation. Therefore, according to Mark Knickrehm, organizations need to begin preparing employees to work alongside smart machines and add value to their efforts.

Implementing AI raises ethical questions.  Other articles in the book, including one by Roman Yampolskiy, suggest that it’s not too early to consider the ethical concerns around AI. Algorithmic bias and lack of transparency are two critical issues that AI exacerbates. These powerful technologies have powerful implications for the workplace and the broader society.

With these fundamentals covered, it’s time to dive into the articles. To best understand how AI will impact your company’s situation, consider these questions as you read:

• Which particular AI technologies have the greatest potential benefit to your organization?
• How might those technologies enable new strategies, business models, or business process designs?
• What data resources do you have—or might you obtain—in order to power AI projects?
• How do you anticipate that AI will impact your workforce, and how can you begin to prepare employees to augment AI capabilities?

If you and your organization haven’t already confronted these questions, let this book spark conversations. Think about how the right AI initiative could help your division perform better or make you more efficient at your own job. Simply asking the questions may be the first step in starting your company down the path of transformation.

As a professor and a consultant on information technology and business, I’ve spent the past several decades watching AI alternate from spring blooms to winter doldrums. This time is different. AI is deeply ensconced in business and is starting to bring about exciting change. Now, it appears that winter will not return.