The enterprise AI conversation has changed — and if your organization hasn't noticed, your budget is about to make it painfully obvious.
A few quarters ago, the executive question was "Are we using AI?" Today it has shifted to something sharper: "Can we prove AI is creating leverage?" And according to a new first-party report of 421 executive responses published by Open Future Forum in July 2026, more than half of enterprises have already put their wallets on hold until someone can answer that question with real numbers.
This is not a slowdown in AI adoption. It is a maturation. The executives surveyed are past the exploration stage — 71% at a major 2026 finance event said their teams were already running Claude or another AI tool. Only 8% hadn't started at all. The delay in spending has nothing to do with hesitation about AI's potential. It has everything to do with a hard-won lesson: enthusiasm does not pay for the next AI initiative. Results do.
The 6-Month ROI Window You Didn't Know You Had
Here is the number that should be on every AI lead's radar: 62% of finance-room executives expect measurable return on an AI investment in under six months. Seventy-nine percent expect it within a year.
Those expectations create a specific and unforgiving clock.
Most enterprise AI programs are not designed with this in mind. The conventional wisdom — build the foundation, scale carefully, measure later — runs headlong into a board room that wants evidence at the next quarterly review. If you are running an AI program that will take 18 months to prove value, you are building for a buyer who wants proof in six. That is not a timing inconvenience. It is a structural risk.
To make it worse: Gartner's 2026 data shows that only 41% of AI agent rollouts cross positive ROI within 12 months. Nineteen percent never reach payback. That means the average enterprise AI initiative is already in tension with what executives expect — and nearly one in five will never close the gap.
The implication is clear: ROI measurement is not something you bolt on after the pilot. It is something you design before you start. Every AI initiative needs a baseline, a measurement framework, and a 90-day checkpoint built in from day one. Not because the board asked for it. Because the timeline they expect is shorter than most programs are built to deliver.
The Main Blocker Is Not What You Think
You might expect the biggest obstacle to more enterprise AI spend to be security concerns, integration complexity, or talent shortages. All real challenges. All frequently cited.
But the Open Future Forum data tells a different story. Among finance-room executives, 53% named proving ROI as the main thing stopping them from spending more on AI — far ahead of any other single blocker.
Executives are not rejecting AI. They are rejecting unmeasured AI.
That distinction matters enormously for anyone inside an enterprise AI function. The problem is not skepticism about the technology. The problem is that too many AI programs have delivered activity without evidence. They have produced demos, pilots, and dashboards full of engagement metrics — but not the cost savings, revenue attribution, or productivity benchmarks that stand up to a CFO's scrutiny.
The market has moved from excitement to evidence. The next dollar of AI investment is gated by proof of the previous dollar.
Talking to peers across the industry, I keep hearing the same pattern. The AI program got funded on a wave of optimism. A year later, the CIO is standing in front of the CFO trying to justify the renewal — and the data isn't there. Not because the AI didn't work, but because no one defined what "working" meant in financial terms before they started.
Budget Substitution: The Signal Nobody Wants to Say Aloud
Among finance-room respondents in the Open Future Forum report, 17% said this year's AI budget comes at least partly from money that would otherwise have gone to headcount.
Roughly one in six.
This is the number that tends to make people uncomfortable in all-hands meetings — but it deserves to be discussed honestly at the leadership level. AI is no longer being evaluated purely as a software investment. For a meaningful minority of finance leaders, it is being evaluated against hiring.
That is not a statement about layoffs. Most organizations in this cohort are still growing. But it signals something important: the ROI conversation has moved beyond "AI saves our analysts time" to "AI changes how we think about capacity planning."
When a VP of Finance tells me that a team of three can now do what previously required five — and they are choosing not to backfill those two positions — that is a real-world expression of the budget substitution trend. AI is starting to change the unit economics of knowledge work at the enterprise level.
For business leaders, this is a strategic opportunity. For technical leaders, it is a mandate to build AI that is measurable enough to justify that kind of decision.
Who Signs vs. Who Proves: The Executive Authority Gap
This is where the data gets particularly instructive.
When asked who signs off on new AI purchases: 47% named the CEO, 26% named the CFO or Finance, 21% named the CIO or CTO.
It sounds like the CEO owns AI. In the purchase decision, they do. But buying a technology is not the same as proving it worked. And in enterprise AI, the person who holds the pen on the purchase is often not the person who has to stand behind the results.
The CFO is increasingly on the hook for proving AI ROI — even when they weren't the one who championed the buy. Finance has to produce the numbers, reconcile the cost lines, and translate vague productivity claims into the language of auditable business outcomes. When the CEO says "we are all in on AI," it is Finance that has to figure out what that looks like in the model.
This gap — between who approves and who proves — explains a lot of the friction inside enterprise AI programs. It explains why AI vendors are being pushed harder to show usage data, outcome metrics, and business value dashboards. It explains why CIOs are building ROI frameworks they never needed before. And it explains why CFOs have become central figures in AI governance conversations that used to be IT-only discussions.
For AI leads inside enterprises, this has a practical implication: do not build your success story for the CEO who bought. Build it for the CFO who has to defend it.
The Seat Gap: Vendors Are Selling to the Wrong Buyer
The Open Future Forum report surfaces one more insight that matters for enterprises evaluating vendor relationships.
Among AI founders charging for their products, 50% price on usage, 18% on outcomes, and 25% on per-seat licenses. The shift toward usage and outcome pricing is real and meaningful — those models are more auditable, more aligned with value, and more defensible to a Finance function.
But here is the irony: when asked who they sell to, AI founders most often name the CIO/CTO and business unit leaders. Finance remains less visible to them as a buyer.
This is what the report calls the seat gap. Inside enterprises, the CEO and CFO have become central to AI approval and accountability. But vendors are still building their sales motions around the CIO and the business unit. The people who now hold the purse strings — and the proof burden — are being underserved by the very vendors trying to get into their stack.
For enterprise buyers, this creates a specific kind of leverage. Vendors are more willing than ever to prove outcomes before you buy, precisely because the market has taught them that CIOs can no longer make standalone decisions. Use that leverage. Demand outcome-based pilots. Require baseline measurement as a condition of engagement. Ask vendors to show you how they will prove ROI to your CFO — not just how the technology works.
What CIOs and CTOs Need to Do Now
The signal in this data is clear for technical leaders: your job has expanded. You are no longer just the person who selects and deploys AI systems. You are now the person who has to make those systems legible to Finance.
That means three things in practice:
First, define ROI before you deploy, not after. Every AI initiative needs a baseline. Time-per-task, cost-per-transaction, error rates, headcount allocation. Without a before, there is no after. Without an after, there is no renewal.
Second, build for the 6-month proof window, not the 18-month roadmap. If your AI program's first meaningful result comes at month 14, it will face a budget review at month 6 with nothing to show. Structure your roadmap so that something measurable — even if narrow — is visible before the first annual review.
Third, own the CFO relationship. The data is clear: Finance is now a primary stakeholder in AI governance. If the CFO in your organization only hears about AI through cost line items and renewal requests, that is a gap you need to close. Proactive briefings, ROI dashboards, and regular updates on AI value attribution are no longer nice-to-haves. They are survival skills for enterprise AI programs.
What CFOs and Business Leaders Need to Ask
If you are a CFO or business unit leader who has signed off on AI spend in the last 12-18 months, the Open Future Forum data raises three questions worth putting on the table:
Do you have a six-month ROI checkpoint? If your AI program does not have a defined measurement milestone within the first half-year, you are operating on hope rather than evidence. That is fine as a starting position — but not as a steady state.
Who is accountable for proving the return? The authority gap is real. If the CEO bought it and the business unit is using it, but nobody owns the outcome reporting, you are likely sitting on an AI investment that will be difficult to justify at renewal. Accountability needs to be explicit before the invoice arrives.
Are you measuring the right things? Engagement metrics — logins, queries, usage hours — are activity, not value. CFOs need to push for financial proxies: cost avoidance, error reduction, cycle time, revenue per head. Those are the numbers that hold up in a budget review.
The Bottom Line
The Enterprise AI Leverage Report's headline finding is not just that enterprises are being more rigorous about ROI. It is that the window for that rigor has compressed dramatically. Sixty-two percent of executives expect measurable returns within six months. Half of all AI spend is blocked pending proof.
That is a very different environment than the one that existed 18 months ago, when AI pilots were bought on enthusiasm and given indefinite runway to mature.
The market has moved from excitement to evidence. Every enterprise AI team — whether you are building, buying, or governing AI — needs to operate accordingly.
The executives who figure this out will keep their AI budgets intact. The ones who don't will find themselves explaining, at the next board review, why the money produced activity but not results.
Six months. That is the clock you are working against.
Sources: Executive AI Leverage Report, Open Future Forum (July 2026) — 421 executive responses across finance, security, growth, and founder rooms. Gartner 2026 AI agent ROI data via samta.ai.
