For today’s CFOs, accurately understanding the cost-benefit of AI is growing more important as the technology is woven more deeply in their companies.
CFOs are not just asking what the return on investment is from the technology and what AI use cases are most effective, but “just because we can do it in AI, should we do it in AI?” Board CFO Gordon Pothier said.
“I think there's a first level of us really just trying to understand, [get] our arms around the cost” of the technology, Pothier told CFO Dive in an interview focused on how most finance chiefs approach the AI-ROI question. “How can we plan around this?”
Healthy skepticism
Generative AI has swiftly become a standard part of the toolkit among many professionals and business processes since its debut in 2022 with the launch of OpenAI’s ChatGPT.
For example, Board’s engineering team was an early adaptor, and “I think, right now, if we tried to take away these tools, there would be a complete uprising,” Pothier said.
Pothier took the top finance seat for the Boston, Ma.-based enterprise planning platform in November 2025, according to a company press release. Prior to Board, he served a three-year span at code security provider Sonar, including a two-year term as its CFO. His past CFO roles have included serving as finance chief for Onapsis Inc. and Carbon Black.
As more employees begin to rely further on AI, however, CFOs are paying closer attention to its price tag — and to the potential risks of over-dependency, both in terms of costs and effectiveness.
Last month, for instance, Uber executives revealed the rideshare company had blown through its 2026 AI budget in just four months. The reports triggered rising concerns from technology and business executives, many of which need to balance the growing cost of AI tokens — the units of data processed by AI, purchased by businesses using the models — against its still-to-come ROI.
When it comes to AI projects and spending today, “there's a little more healthy skepticism,” among CFOs and other key company leaders, Pothier said. There’s still a push to adopt the technology — Gartner anticipates AI spending will top $2.5 trillion by the end of this year, according to a May report — and there are still valuable use cases for generative AI, he said.
But, “I do think that skepticism is starting to come in in a healthy way around, ‘okay, are those cost savings real in a scalable way or are they short term?’” Pothier said. “What is that ROI? And now [we’re] starting to really talk more about, what's that cost look like in in the years to come?”
Filtering out the noise
When it comes to AI over-dependency, Pothier said the danger for companies may come when they “are jumping forward too fast for whatever reason, and not thinking about all sides of the equation.”
For example, when it comes to token usage, businesses may be calculating what tokens cost them now, but “how do I scenario plan for what that looks like in three to five years?” he said.
For CFOs, that means more scrutiny when it comes to both internal AI experimentation and when choosing third-party providers that are truly sustainable, Pothier said.
For example, as many businesses begin to tap AI to improve their marketing efforts, failing to do so may seem like falling behind — but it’s important for CFOs to “really question” what the AI tool on offer is actively doing and how it works, he said.
“Is a chat bot that is not really doing much more than any other chat bot?” he said. “What's the token usage again there too? Is it processing … using tokens in a way that's going to be really, really expensive?”
That’s not to say CFOs should avoid integrating AI wholesale — if there is a solution that can solve a particular problem, “you should absolutely look at it,” Pothier said. “But I think somehow you have to try to filter out some of the noise and prioritize, and that's not always easy.”
