Coupa Surveyed 600 Finance Leaders. 41% Say Autonomous AI Agents Will Drive the Biggest Returns. Only 13% Have Deployed One Successfully.

October 27, 2026 · Josh Agusti, Founder, Aiven Consulting Group

Ask CFOs where AI's financial return will actually come from, and the answer has shifted. Coupa's 2026 Strategic CFO Report surveyed 600 finance leaders worldwide, and 41 percent said autonomous AI agents that execute workflows on their own, not tools that merely summarize or report on data, will deliver the greatest long-term financial return.

That marks a real break from how most finance functions have used AI so far, which has leaned heavily toward dashboards and forecasts rather than systems that act. The harder part is what happens after the belief is stated. Only 13 percent of those same CFOs say they have actually deployed agentic AI successfully.

Why execution, not insight, is where CFOs now expect value

The shift matters because of what finance leaders are implicitly admitting. Most finance organizations already have AI that drills into variance reports, forecasts cash flow, or flags anomalies for a human to review. What 41 percent of CFOs are now betting on is different: software that initiates the close process, routes an exception to the right approver, or reconciles a mismatched invoice without anyone opening a ticket. Coupa's own release frames the shift as a move away from data summarization and toward autonomous workflow execution, and that framing fits a broader pattern across finance organizations that have already automated the easy reporting layer and are now looking for where the next unit of value sits. For a finance function evaluating which use cases are worth building first, a working agent library of vetted, auditable workflows is a more useful starting point than a long list of theoretical capabilities.

Despite the appetite, 85 percent of the CFOs surveyed said AI is now central to their financial strategy, and all of them plan new AI investment in the next six to twelve months, so the 13 percent success rate is not a funding problem. That pattern is not unique to finance. Gartner expects well over 40 percent of agentic AI projects to be canceled by 2027, and the reasons rarely trace back to the underlying model. A closer look at why agentic AI projects fail shows the same culprits CFOs described here: unclear data ownership and brittle integrations between the systems an agent needs to touch.

Why fragmented data, not budget, is the real blocker

Asked what is actually standing between ambition and a working deployment, CFOs in Coupa's survey pointed most often to data fragmentation, not budget, staffing, or model quality. An autonomous agent that reconciles an invoice or routes an approval has to pull from the general ledger, the procurement system, the bank feed, and often a separate contract repository, and if those systems don't agree on what a vendor record or a cost center actually is, the agent either fails quietly or escalates everything back to a human, which defeats the purpose of deploying it in the first place.

For a CFO weighing whether to greenlight an agentic finance pilot, the more useful question is not whether agents can eventually pay for themselves, most now assume they can. It is whether the underlying data across the ledger, procurement, and banking systems is clean and connected enough for an agent to act on without creating a new category of reconciliation work. Finance teams that want a clearer picture before committing budget can use an AI ROI calculator to model where the return is most likely to show up first, informed by the obstacle Coupa's CFOs actually named.

Key Takeaways

Coupa's 2026 Strategic CFO Report found that 41% of the 600 finance leaders surveyed now believe autonomous AI agents, not reporting or summarization tools, will produce the greatest long-term financial return from AI.
Despite that confidence, only 13% of those same CFOs say they have successfully deployed agentic AI, even though 85% call AI central to their financial strategy and all of them plan new AI investment within the next year.
The leading obstacle CFOs named was not budget or technology but fragmented data across ledger, procurement, and banking systems, the same disconnect that causes many agentic AI projects to stall or get canceled elsewhere in the enterprise.