Book interview: ‘All in on AI’ by Thomas Davenport

When it comes to recognising what artificial intelligence can do for their businesses "most companies are just scratching the surface", says Thomas H Davenport, co-author of 'All in on AI'. "They have a few pilots or proofs of concept under way that don't really have any substantial impact on their businesses," he continues. "But in our book, we describe the extreme adopters: those who are doing many things with AI, some of which have the potential to transform their strategies, business models or key business processes."

Observing these "highly aggressive" companies will give other, more reluctant adopters, ideas about how they should proceed with AI, he argues.

Subtitled 'How Smart Companies Win Big with Artificial Intelligence', the book's message could not be clearer. Whichever way you interpret the deliberately ambiguous word 'smart' – either in the sense of intellectually agile, or progressively digitally enabled – if you're not one of Davenport's aggressive adopters, you're losing out.

"We wrote the book in order to educate other companies about what it takes to succeed with an 'all in on AI' approach," he says. There's another ambiguity, too: while 'all in' can be taken to mean 'total commitment' as a calculated strategy, it carries a more specific connotation from poker, where the term refers to the maximum 'raise' of betting your full stack of chips. Perhaps there's an implicit gamble in being an early AI adopter, but the book is full of examples of people who've won big.

In a publishing market currently awash with titles on AI self-help/inspiration, it's tempting to wonder what 'All in on AI' delivers that others don't. Davenport leaps to his (and co-author Nitin Mittal's) defence, stating: "Perhaps the most distinctive aspects of the book are the set of factors that distinguish highly aggressive adopters of AI, and then the examples of how several companies are employing each factor."

He continues that "one of the most interesting to me is the adoption of new business models based on ecosystems. We describe how AI supports a set of five major ecosystems at Ping An, the largest private-sector company in China. We also discuss an ecosystem called SkyWise at Airbus, which comprises Airbus itself, all the airlines that fly Airbus aircraft, and the planes themselves. AI is used to integrate the data, predict maintenance needs, and optimise routes and fuel consumption."

One of the big revelations of the book is that only 1 per cent of big companies have 'all in' integrated AI strategies. It seems that organisations are currently only able to get fully into AI if they have a relatively large number of low-risk areas in which to apply it.

"Some, like DBS Bank and Anthem [now Elevance Health] had some early failures with AI that wasn't yet mature technically. But now they are putting in place a variety of lower-risk applications that together have a strategic impact." He cites Capital One as an example of an organisation that "may have thousands of AI applications, but each one is pretty straightforward to develop". Clearly, going 'all in' requires heavy financial commitment to integration, if it is to pay off, and Davenport says "we only profiled companies that are already receiving substantial benefits from those investments".

When Davenport talks about "aggression" in the context of technology adoption, there are two key factors. First is early adoption, and second is funding. Time and again, 'All in on AI' makes clear that there aren't many gains to be made by the faint-hearted. This advocacy for decisiveness is something that the visiting professor at the University of Oxford has written about before. Way back in 2007, Davenport's book 'Competing on Analytics' described how companies that made 'aggressive' use of Big Data could realise a competitive edge. The book became a bestseller and was popular "even when companies didn't want to adopt analytics so aggressively", if only because they found it useful to learn what their competition was doing. Likewise with AI: "While you may not want to adopt all the leading-edge practices, you need to consider them. If the companies we profile are in your industry, it may be necessary to at least keep up with them in order to survive."

'The losers will be those that only experiment with AI on a small scale'

Before organisations make the transition from passive observer to active participant they need to have a strategy in place, in order to coordinate "all the technical, business and organisational change necessary to succeed with large-scale AI". Davenport says that while some companies have a separate AI strategy, or have combined AI with a broader digital strategy, "more have what might be called 'an AI-intensive business strategy'. The AI components of it are fully aligned and integrated with the desired changes in business strategy, models and processes. That's why the CEO and the senior management team need to be aligned with it."

Given worries about AI's effect on human jobs, Davenport notes that "we didn't find a single example of large-scale job loss from AI or automation in any of the companies we researched for the book. This could change once we get highly competent general AI. But that time isn't here yet."

Davenport does say he was surprised by the wide global distribution of AI adoption among big companies. One flagship adopter is China's Ping An (see extract), "and one of the world's leading banks in the use of AI is DBS Bank, based in Singapore. In addition to a number of US firms, we also found others in the UK, Europe, Canada and Japan. AI is a global phenomenon."

He believes that the early adopters will be the big winners. "Some already are pulling away from their competitors. We saw similar patterns with earlier technologies: Walmart, for example, became dominant in retailing because of its supply chain technologies, satellite networks and RetailLink connections with suppliers. The winners today will be those companies that adopt AI early and aggressively and integrate it well with their businesses. The losers will be those that sit on the sidelines or perhaps only experiment with AI on a small scale." N

'All in on AI' by Thomas H Davenport and Nitin Mittal, is from the Harvard Business Review Press, £25.