Call it the generative AI paradox.
CFOs are chasing the hope of a bonanza from generative artificial intelligence in one of the most frenzied investments in technology in decades.
Yet they have, at best, a loose grip on how to measure the innovation’s ultimate business payoff, according to financial executives and AI experts.
“I haven’t seen anything quite like this,” said Daniel Rock, who co-founded Workhelix, a generative AI consultancy, with Erik Brynjolfsson and Andy McAfee.
Since the public launch of ChatGPT in November 2022, “there has been this ‘Ah-ha!’ moment, when people say, ‘This is capable of things that I didn’t think information technology systems could do,’” Rock said in an interview, noting the prospect that AI will streamline operations, better serve customers, upgrade forecasting, speed software programming and improve marketing and sales.
The CFO rush to generative AI will likely prove especially rapid because adoption is happening primarily from the bottom-up — from consumers — rather than top-down — from businesses, San Francisco Federal Reserve Bank President Mary Daly said.
The release of ChatGPT marked a “surge moment,” as “generative AI went from a thing others did to relative ubiquity overnight,” Daly said at a University of Chicago conference in March. “This moment led to what I think of as ‘people-driven diffusion.’”
As a result, CFOs and their C-suite colleagues face shareholder pressure and rising expectations for AI-enabled efficiencies and customer service, Daly said. They must move “from apathy to action” faster than with other formative innovations including steam power, electricity and the internet.
“Our conclusion is that you will essentially see the benefits of generative AI over the next five years,” EY-Parthenon Chief Economist Gregory Daco said in an interview. To hit a similar payoff, computers took 10 years, electricity more than 40 years and steam power 80 years, Daco said.
Rapid adoption of generative AI will enable companies to better innovate, harness the new skills of employees, retain customers and gain market share, Daly said.
At the same time, quick deployment of generative AI requires CFOs to measure ROI faster than prior technologies or risk wasteful spending, according to financial executives and AI experts.
AI shakedown
CFOs must also look out for hype and outright fraud, according to the top cop at the Securities and Exchange Commission.
“While perhaps not quite yet a perfect storm, there’s certainly one brewing around AI,” SEC Enforcement Director Gurbir Grewal said. “Every day, we see revolutionary technological advancements in artificial intelligence, or AI, that promise to transform nearly every aspect of our lives, including our financial decision making.”
“Every day, we see individuals, corporations, analysts and others touting these developments,” he said in an April speech. “Every day, we see companies attempting to not only develop AI capabilities, or harness it to improve their productivity and growth, but also to attract and retain investors.”
Most investors (61%) believe that rapid adoption of AI is very or extremely important to a company’s value, according to a PwC survey.
In turn, CFOs and their C-suite colleagues harbor “fear of missing out,” Glenn Hopper, CFO at Eventus Advisory Group, said in an interview. “They’re getting pressure from investors and boards — at every minute, from everybody.”
Consumers have also set a high bar, according to financial executives and AI experts.
“At some companies AI deployments will be something that must be done because customers expect it,” according to Rock, who is also assistant professor of operations, information and decisions at the University of Pennsylvania’s Wharton School. “It’s just going to be table stakes.”
During the coming year, 37% of U.S. companies plan to invest at least $100 million in generative AI, KPMG found in a survey of 220 companies published in March.
Spending on AI software will surge to $298 billion by 2027, 140% more than in 2022, according to Gartner, with 35% of investment focused on generative AI.
“You’re not going to be able to stop the momentum,” Hopper said.
Nvidia, arguably the decade’s top momentum stock, leads the juggernaut behind generative AI adoption. Its share price during the past year has rocketed 153%.
Yet a hint of speculation by Nvidia on the operational value of its products suggests that the onrush into the new technology may, at some point, venture onto shaky ground.
Countries as well as companies have hired the company to upgrade $1 trillion worth of data centers into “AI factories to produce a new commodity — artificial intelligence,” Nvidia CEO Jensen Huang said in a quarterly earnings call on May 22. “The next industrial revolution has begun.”
Every $1 invested by a cloud service provider in Nvidia’s software and networking can, over four years, yield $5 in revenue, Nvidia said in a “Company Overview” on a page titled, “The High ROI of High Compute Performance.”
In a footnote, however, Nvidia said in the May report that its calculation is merely an “illustrative example.”
To be sure, CFOs gauging ROI in generative AI are trying to map fuzzy terrain, the financial executives and AI experts said. Some payoffs from the emerging technology are potentially illusory, especially in the short term.
Only 5.2% of organizations attributed more than 10% of their earnings before interest and taxes to use of generative AI, McKinsey said in May, citing a global survey of 1,363 executives.
Big Data, Big Challenge
The challenge of measuring the value of data predates generative AI, as well as the advent of so-called Big Data and advanced analytics two decades ago.
CFOs and investors have struggled for years to precisely value proprietary data on customers, markets, products and other information.
The measurement challenge did not stop the meteoric rise of Amazon, Google, Uber and other Silicon Valley giants that rely on detailed, self-generated data to fuel profits and growth.
The path-breaking companies rode an explosion in the use of the internet, smartphones and other tools for connectivity and data sharing. Investors place more value on their proprietary data than on their buildings, workers or technologies, Laura Veldkamp, a finance professor at Columbia University’s Business School, said at a New York Fed conference on productivity in February.
CFOs in industries that arose before the advent of so-called information technologies often face a bigger challenge gauging a return from generative AI across their operations, Rock said. “I’m not sure that ROI in gross terms is going to be there.”
AI cost estimates by some companies have fallen short of the mark by as much as 1,000%, Nisha Bhandare, a Gartner vice president and analyst, said at a Gartner CFO conference in May. “Given how new AI is, we don’t really know how much it costs, and are learning as we go.”
Editor’s note: This is the first of two reports on the challenge of measuring return on investment from generative AI. In part two, CFO Dive will describe how financial executives can limit the cost and maximize the payoff from the technology even though, at the outset, they may lack a solid estimate of the potential gains.