CIOs are looking for ways to manage AI’s costs and proliferation — and they’re not alone.
Nearly all enterprise FinOps teams, which emerged as a cloud management discipline, now manage some aspect of AI spending, with outcomes and ROI increasingly replacing cost savings and optimization as a measurement of success, according to Jay Litkey, SVP of cloud and FinOps at software company Flexera.
“Value delivered to the business is now used by about 64% of teams as a KPI,” Litkey said during Flexera’s FinOps Forward 2026 virtual event on Wednesday.
But these teams are poised to do more, helping tackle tasks such as establishing agent inventories and standards to help enterprises — and CIOs, specifically — manage potential AI sprawl and the onslaught of shadow AI. Nearly 4 in 5 FinOps teams report to the CIO, according to the FinOps Foundation’s State of FinOps 2026 report.
FinOps teams don’t police how enterprises are investing in and using AI, but help with the governance requirements and metrics necessary to determine what effective AI use looks like across an organization, said panelist Salomé Keet, head of software asset management and FinOps at FirstRand, a financial services firm based in South Africa.
For FirstRand, proving the business value of the bank’s IT spend on projects, including AI, involves more than just FinOps. Setting the course to value means breaking down silos and working together with teams across the enterprise, she added.
“In terms of spend, that’s something we are tracking very closely with finance, with procurement,” she said during the event. “Pulling all of the data sources together is a big focus for us to make sure that, while everybody is so excited about this and while this is a growing phenomenon, we make sure that we keep tabs and make sure that we’re not duplicating.”
Keet said she’s also looking at how AI can be used more effectively within the FinOps team itself, which she said could be a gamechanger for their operations.
Managing AI
FinOps teams face three major challenges when it comes to managing AI spend: understanding its full scope, quantifying value and equitably allocating the cost, said Steve Trask, COO of the FinOps Foundation, a project focused on technology value management, during the event.
“It’s definitely a hot topic within the community and how we build better practices around this,” Trask said of AI spend.
Only 7.5% of enterprises are embedding FinOps into AI projects, which makes it difficult for teams to stay on top of spend, said Jevin Jensen, research VP, intelligent CloudOps market service at IDC, at the event. As companies move forward on AI without FinOps baked into the projects, 41% of enterprises are wasting more than 15% of AI spend, he said.
But FinOps teams could do more than manage spend. AI agents, for example, are still relatively new but will soon present an additional challenge to enterprises. As the number of agents begins to scale, FinOps teams could help create standards, protocols and inventories to tackle uncontrolled proliferation.
Enterprises deployed 28.8 million agents in 2025, Jensen said. By the end of 2026, IDC expects enterprises to be managing 80 times that number. Daily agent actions will near 217 billion by 2029, he added. As more agents are built and embedded in mission critical systems, enterprises might struggle to monitor AI deployments, fueling shadow IT growth, Jensen said.
“I’m worried that shadow IT is coming back with a vengeance when it comes to these AI agents and how easy they’re going to be to create,” he said.
