After several years of record levels of investment, enterprise AI budgets are still growing. According to our latest enterprise technology spending survey, 98% of respondents are increasing their AI spend over the next 12 months, and not a single CXO surveyed is cutting back. But as budgets scale, it’s putting pressure on a harder question: Where are the returns?

Most organizations are finding some measurable returns for AI—but very few can prove how meaningful AI is to the overall business. Fifty-three percent of enterprise technology leaders say they’re seeing clear ROI from some area(s) of AI investment today. But only 6% have a well-defined measurement framework tracking ROI consistently across the organization. Another 42% are measuring in some areas but not consistently, while 43% are still defining how to measure at all.

That gap—between organizations that believe AI is working and organizations that can prove it’s working, with demonstrable ROI—is one of the most consequential issues in enterprise technology right now.

Most projects are not yet delivering measurable ROI

Fourteen percent of organizations report that none of their AI projects have delivered measurable ROI yet and 31% say fewer than 25% of projects have done so. Only 16% of survey respondents reported that more than half of projects are delivering measurable returns.

That doesn’t necessarily mean the investments are wrong or misguided; it may mean measurement infrastructure hasn’t kept pace with deployment velocity. And when leadership asks where the money is going, most technology teams don’t yet have a complete answer.

Where the returns are clearest

Twenty-eight percent of CXOs surveyed named software development and engineering productivity as the area of clearest ROI within their organization. These functions, which comprise coding assistants and automated QA testing, are delivering measurable productivity gains that are relatively easy to track. Internal operations / workflow automation and customer support round out the top three.

How ROI is being measured

The metrics skew toward cost and efficiency. Cost savings or operational cost reduction is the most common measurement approach at 76%, followed by employee productivity or time saved at 73%, and speed of delivery at 61%. Revenue-based metrics lag behind—only 38% are tracking revenue generated or attributed to AI—reflecting both the difficulty of attribution and the earlier-stage nature of revenue-generating AI deployments.

That skew shows up again when we asked leaders where they expect the value to eventually land. Forty-six percent of organizations expect AI value to accrue as a balanced mix of cost reduction and revenue growth. Twenty-six percent expect it to hit primarily the bottom line, while only 16% are banking primarily on top-line growth.

The budget pressure is building

Nearly 80% of AI budgets have some reallocation component—pulled from existing SaaS spend, headcount or infrastructure. As that reallocation becomes more visible in budget reviews, the ROI conversation will intensify. Organizations that build consistent measurement frameworks now, before the scrutiny arrives, will be in a far stronger position when the questions get harder.

The ROI question is no longer a future problem. It’s the present one.

This post is part of Battery’s 2026 Enterprise Spend Report series. Download the full report.