Industry leaders show the wisdom of going all-in on AI

Despite all the attention generative AI is getting right now, most organizations have done little with artificial intelligence. That is a big mistake, says Tom Davenport, senior advisor to Deloitte’s Analytics practice. Enterprises, especially industry leaders, need to be all-in on AI if they are to remain competitive.

To truly benefit from AI investments, organizations must rethink how humans and machines interact in work environments, Davenport says, starting with applications that change how employees do their jobs and how they interact with customers. AI should help drive every business decision, and every product or service offering.

That message is at the heart of Davenport’s new book, All in on AI: How Smart Companies Win Big with Artificial Intelligence, co-written with Deloitte Consulting principal Nitin Mittal.

“It doesn’t provide a lot of value to just tinker with AI — to do an experiment here and there,” Davenport says. “We can do AI pretty easily on a small scale. But integrating it into how you do your work means embedding it into your existing technology architecture.”

Instead, organizations need to incorporate AI into business processes and workflows. They need to upskill staff to work with AI. And they need to ensure the AI technology can scale. They also need to do these things over time, to ensure the world doesn’t change in a way that makes the effort counter-productive.

“There are lots of benefits to be had from aggressively adopting [AI], and using it to change your strategy, your business model, and your key business processes. It’s really urging companies to do more with AI than the toes-in-the-water approach that most have taken,” Davenport stresses.

Learning from those who reap huge benefits from AI investments

In the book, Davenport and Mittal identify 30 organizations that have gone ‘all-in’ on AI and benefitted enormously from this strategy.

“The most impressive example is Ping in China,” Davenport explains. “Most people don’t know much about it, although it’s the 16th largest company in terms of revenue in the world. It is now the largest private sector company in China. Ping was founded as an insurance company, but they now have five ‘ecosystems,’ or business units. In addition to insurance, they have added banking, healthcare, Smart City [a smart cities business], and an automobile services business.”

Ping has grown at an unbelievable rate, Davenport says. The firm created this ecosystem approach to enable them to partner with other organizations and get customer data from those relationships. They use that data to create AI models that do a good job of predicting or categorizing behaviors. They then grow each business and get more data, he says.

“My favorite example is their healthcare business, which created an offering called Good Doctor,” Davenport continues. “During the pandemic, we were impressed in the United States when people could talk to their doctor over Zoom, get a prescription or whatever. But this goes so far beyond that.”

Good Doctor is an AI-based system for triage, diagnosis, and then treatment recommendations, Davenport explains. An actual doctor makes the final diagnosis and recommends treatment for a patient, but the physician gets recommendations from the Good Doctor system.

“For me, the most astounding thing is that it is used by nearly 400 million people in China, more than the population of the United States. They don’t have enough doctors in China, so it has made a huge difference to the state of healthcare,” Davenport says.

Other AI leaders transforming their markets

Another example profiled in the book is Shell Oil Co., which has embraced AI for many of its business units, and has used AI to re-engineer a number of processes. The most dramatic example is around inspection of Shell’s large plants and pipelines.

“It used to literally take up to six years to inspect every aspect of their plants with human inspectors,” Davenport says. “Shell now shoots for six days, using drones and AI-based image analysis systems. They have achieved dramatic reductions in the time to do these inspections, and there is a potential safety benefit here as well. Shell has also trained over 5,000 engineers to be citizen data scientists, in a sense. They are able to interpret this inspection data without having a professional data science background.”

A third example is Kroger, which is one of the largest grocery retailers in the US. Kroger has a wholly-owned data science subsidiary called 84 Point 51 Degrees, based in Cincinnati. The name comes from the longitude of Cincinnati.

“The subsidiary is really quite impressive in terms of the data science work that they do for Kroger related to consumer products, as well as the companies that sell their products in Kroger,” Davenport explains. “For example, they run a huge model that predicts the sales in every stock unit, in every store, in their entire collection of stores, every night.”

Kroger also has the largest grocery loyalty program in the country. The company uses data from that program to predict what product offerings and promotions will convince members to show up at a local store more often, and to buy more.

“They are using the loyalty program to recommend new products with a high level of nutrition, to encourage customers to shop in the healthy food space,” Davenport says. “They also sell some of their data insights to consumer products partners. I think they’re well ahead of any other grocery retailer in that regard.”

AI’s value for large legacy organizations

The primary focus of Davenport and Mittal’s book is on legacy organizations that want to truly transform with AI.

“It isn’t about the digital natives that have a much easier time of it, since everybody [in those organizations] already believes in AI and digital transformation,” Davenport notes. “Still, a lot of companies say they’re doing it. But they have very few deployments of AI to show. They haven’t integrated it into their day-to-day work, and hence, don’t get any real economic value.”

Davenport acknowledges that many companies may feel reluctant to make a large investment at this relatively early stage of ‘modern’ AI. But the book is intended to demonstrate how organizations committed to AI use are reaping significant benefits, and in some cases, transforming their markets.

Toward that end, these leading organizations are both broad and deep in terms of AI adoption, Davenport says. They have several use cases or applications in production. They use a variety of technologies, including machine learning. Many also use robotic process automation and linguistics-based computational chatbots.

“The time to stand on the sidelines is over,” Davenport stresses. “In a way, we were trying to scare readers and say, ‘It’s going to be hard to catch up if somebody else in your industry is doing this, and you’re not.’”

Most importantly, AI is an area where it will be difficult to be a fast-follower, because it requires a lot of data and a lot of skills that are not widely available, Davenport explains. Organizations should start investing in AI now, and there are ways to do this fairly easily and inexpensively.

“Many vendors are incorporating AI capabilities into their ERP systems and CRM systems, so you could start there,” Davenport says. “But if you want any sort of competitive advantage from AI, you probably have to develop some of these capabilities yourself. That means developing the skills and technology capabilities in order to produce some of your own use cases.”