AI will yield biggest return in CFO spending during coming decade

By Jim Tyson for CFODIVE.com

• Spending on artificial intelligence that transforms rather than just automates business processes will help CFOs gain a competitive edge more than any other investment in the coming decade, according to Clement Christensen, a director for advisory at Gartner.

• Using AI, companies and government organizations have achieved dramatic efficiencies, Christensen said at Gartner’s CFO & Finance Executive Conference, noting that Amazon increased retail sales 15% to 20%, the Memphis police reduced homicides by 35% and UPS annually saved hundreds of millions of dollars. 

• “AI is not a traditional investment by any means,” Christensen said, adding “there are a lot of unknowns, there are a lot of risks but also incredible potential.” 

CFOs will invest more in information technology this year than ever before, pushing up total spending by 6.2% to $3.92 trillion as they speed up their pre-pandemic plans for digital transformation by at least five years, according to Gartner.

“AI will probably be the most important investment that you’ll make in finance, and maybe the broader enterprise for the next 10 years,” Christensen said. “You have about two to three years to get investing in AI or forever be left behind.”

To date, most spending on AI tends to be ad hoc and focused on automating rather than transforming business processes, he said. For example, a company may use AI to streamline its accounts payable system rather than to identify late-paying clients and flag patterns that can help a sales team avoid seeking prospects who are least likely to pay on time.

“AI should do more than just modernize,” Christensen said. “It should transform your business.”

Currently, only 25% of AI projects move from prototype to production — “a worrying concern for most organizations that are trying to invest in this,” he said. Companies can reduce the failure rate by identifying top business problems and focusing AI projects on finding solutions.

“You cannot forecast ROI for AI, you’ll be making up numbers,” Christensen said. “But what you can do is focus on problems worth solving in your business and align your AI investments to those problems.”

Most companies require 12 to 15 months to cultivate an AI application from inception to production.

“You’ve got two to three years but it’s going to take you 15 months to get started,” Christensen said. “Hopefully, that is a mandate and an alarm signal that you need to get going.”