CFOs Will Lead the Way on AI Cost Efficiency

Information technology officers will be instrumental in deploying artificial intelligence. But CFOs will call the shots on costs and problems.

A new Deloitte study, “State of Generative AI in the Enterprise,” for Q2 2024, is shifting to a more discerning “results” stage with artificial intelligence. The study regularly tracks 2,000 senior executives around the globe, all of whom have either already deployed AI in their workforce or plan to do so in 2024.

In the report, Deloitte concludes that “AI-savvy” companies are “moving past the infatuation stage” with Generative AI and are entering a new phase where they want to “best overcome technical and organizational barriers to create value at scale.”

In this new “what’s the ROI?” stage, chief financial officers will have more say than ever before on how AI is deployed and used inside their companies.

That’s the outlook from Gartner Analyst Nisha Bhandare and Gartner senior director analyst Clement Christensen. The duo provided a clear-eyed outlook for companies building a comprehensive framework to categorize AI initiatives across the enterprise, assess their value, and proactively establish leadership in a new era for organizations. Both analysts spoke at the June 14 Gartner Finance keynote at the company’s CFO & Finance Executive Conference in National Harbor, Maryland.

Noting that CFOs were experiencing “sticker shock” from costly AI implementations, Gartner noted that “the buck stops” with CFOs, who must also embrace a larger role in “directing the organization through its journey in adopting cutting-edge artificial intelligence technology.”

Derailments Ahead

AI deployments can be expected to test the patience of the most seasoned CFO.

“We see real enterprise-level organizational challenges emerging with early [AI] adopters,” Christensen told a crowd of roughly 1,500 people.

Christensen pointed to AI-related “stalls” that could “derail your teams, slow your progress, and could even upend your organizations” if protective guardrails aren’t established. The stalls, Christensen said, include cost overruns, misuse of decision-making, loss of external trust, and mindsets.

Those barriers can affect companies of any size and shape, Bhandare said. “They
require an active engagement…from CFOs,” Bhandare noted. “While all other executives in your organization are focused on AI’s rewards and technical challenges, it will fall on you, the CFO, to navigate the enterprise through these stalls,” she added.

The cost of not addressing those stalls is sky-high, Bhandare noted. Gartner estimates client AO costs were “off” by 500-1,000% with early implementations the data research company examined.

“AI costs go beyond those expected in rolling out the technology,” Bhandare noted. “There are also costs related to usage, data maintenance, and compliance, for example. We’re uncovering these costs with every new AI implementation,” she said.

Fluctuating Costs

There’s good news and bad news for CFOs on that front.

On the upside, AI experimentation and initial rollout phases “start out being very expensive, but then taper off,” the company said in a recent research paper. “Usage costs, however, will become more expensive as more employees start using AI.”

Yet managing these cost fluctuations is no luxury; it’s a necessity.

“It’s this spike in future ongoing costs that your executive colleagues are not planning for in their estimates, and it has the potential of negating all of the ROI of AI as usage continues to grow,” Christensen told the audience.

Avoiding AI-related cost overruns also means CFOs need to avoid blind spots.

Yet Bhandare told the audience that AI only knows whatever its designers have fed it.

“Executives might not think about any blind spots in their knowledge and want to rush to complete automation, putting their firms at risk of AI-caused errors,” she said.

“Finance executives must be the pacesetters of the organization, Bhandare added.


Recent Posts

Leave a Reply

Your email address will not be published. Required fields are marked *