What the AI Revolution Means for Business

• A survey suggests many firms experimented with generative AI in 2023 but few adopted it widely.
• Analysts, executives, and professors argue that there’s more to AI than low-level automation.
• This article is part of “CXO AI Playbook” — straight talk from business leaders on how they’re testing and using AI.

After years of promise and speculation, artificial intelligence for the business world is here, and it’s shaping how companies big and small are planning for the future.

Businesses are increasingly investing in AI, and success stories about companies using AI to cut costs abound. Corporations no longer need to be persuaded to use the technology — but many are still stumped about how precisely AI fits into their operations.

Businesses are interested in AI, but their use of it varies

Most companies are at least thinking about how AI could affect their business. Goldman Sachs strategists found this year that 36% of S&P 500 companies mentioned AI in their fourth-quarter earnings calls. In a PwC survey of more than 4,000 CEOs around the world conducted last fall, 70% said they expected generative AI — a form of AI that creates content such as text and images — to significantly change the way their company creates, delivers, and captures value in the next three years.

“Every publicly traded company in the world today, whether it’s US or UK or Asian markets, is being asked the question by shareholders ‘What is your AI strategy?'” Umesh Sachdev, the CEO of the AI software provider Uniphore, told Business Insider.

Sachdev said that before the release of OpenAI’s ChatGPT in 2022, companies were buying AI as individual solutions for different parts of the business. Companies now are taking a more centralized approach and looking at AI “as a horizontal infrastructure for the whole business,” he said.

Despite all the chatter, deployment remains low. A survey by MIT Technology Review Insights and the Australian telecoms company Telstra of 300 business leaders across Asia-Pacific, the Americas, and Europe found that while 76% of the companies had experimented with generative AI in 2023, only 9% had adopted the technology widely.

A paper from the National Bureau of Economic Research suggests the manufacturing, information-services, and healthcare industries had some of the highest levels of AI adoption in 2017, while construction and retail had some of the lowest.

For some companies, it’s all hype and no action. Gary Gensler, the chair of the Securities and Exchange Commission, recently described companies that only purport to use AI as engaging in “AI washing.” In March, two investment advisors accused of making false and misleading statements about their use of AI agreed to pay six-figure fines.

AI can help workplaces run more efficiently

The MIT Technology Review Insights survey suggested that AI was most commonly being used to automate nonessential tasks.

Felipe Thomaz, an associate professor of marketing at Oxford University’s Saïd Business School, argued that a divide would appear between companies using AI to simply cut costs and those using it to create value.

For companies in the first category, Thomaz said, the approach is “We are doing everything exactly as we did it before, but now we’re doing it with AI.”

“It’s the least imaginative way of using a multibillion-pound technology,” Thomaz said, “but it dominates the industry in practice.”

Citi, for example, told BI in a previous interview that AI is being introduced incrementally at the bank as part of a multi-year process of transformation.

“Let’s get out of the lab and get onto the factory floor. Get everybody using it for small, little things,” Shadman Zafar, Citi’s co-chief information officer, said. “It is about putting it into everything that you do that will have the big accumulated impact.”

AI can open businesses up to new tech possibilities

For the second category, Thomaz pointed to L’Oréal as an example of a non-tech company doing innovative work with AI. The French cosmetics giant developed an augmented-reality tool to let customers virtually try on makeup. It also launched services designed to provide tailored skincare advice at its brands Vichy and La Roche Posay.

“What they get out of that is better product recommendations, a relationship with customers they didn’t have before, and intelligence about the distribution of skin conditions globally, which goes into product development and new products that go to market, intelligence that you wouldn’t have had otherwise,” Thomaz said.

“They remain true to who they are as a company and to their connection to their customers, but actually use the technology in service of their role and benefiting their relationships.”

Deborah Golden, Deloitte’s US chief innovation officer, said AI had massive potential. But, like Thomaz, she argued that proper innovation would require deploying AI to make day-to-day operations more efficient.

“I feel like we’ve overpivoted a bit on GenAI,” she said. “It’s 1/100th of what AI can truly do.”

She said the technology could open up opportunities for industry “cross-pollination.”

“Historically, we’ve seen industries operating in very siloed fashions,” she said. “What AI has enabled us to do is really look and say how can one industry’s data help us solve another industry’s data problems?”

For instance, Golden said AI technology developed by the aerospace industry is being used by the agricultural sector, as satellite data from space can help farmers manage land changes and decide how to plant crops.

Early adoption of AI will require ongoing support

Sam Berndt, a senior director in Gartner’s supply-chain research organization, told BI that factors like employee resistance could hinder companies’ efforts to scale up AI initiatives.

He said that making AI work for companies would require continuous investment, not just a one-off upgrade.

“Over time, models become less accurate to the kind of real-world situations they’re trying to generate content around,” he said, adding that companies would need to set aside resources to update the underlying data or the models.

Liability and ethical deployment are also considerations, especially as legislation and litigation catch up. In February, Air Canada was taken to court after its chatbot inaccurately told a customer he could claim a bereavement discount after traveling. The court ruled that the airline was responsible for the chatbot’s misleading information and that the company had to compensate the customer.

Meghan Higgins, a senior associate at the London law firm Pinsent Masons who specializes in technology and digital regulation, said the case was a reminder for businesses to take into account the risks associated with AI. She highlighted AI’s risk of hallucinating, or inventing information.

Higgins said that in the future, companies could face defamation cases if AI were to make up things about people, as well as cases about biased outputs and automated decision-making. Such legal cases could lead to more caution around AI.

Despite the risks, there’s never been more enthusiasm about AI’s potential and possibilities. Thomaz said that companies should approach AI as an evolution rather than a single transformation.

“There are things that six months ago were said to be impossible, and today they are imminent,” Thomaz said. “We are no longer in a world of regular change.”