AI Adoption in Product Management Hits 94%, but ROI Lags

Everyone is using it. Few can say what it's for.

A cluster of 2026 surveys agrees on one point: artificial intelligence is now a fixture of the product manager's working day. They disagree on how far that has gone, and on whether organizations are getting anything strategic out of it.

The most aggressive numbers come from Productboard, which surveyed 379 enterprise product professionals. Every team in its sample used AI tools, 96% reported using them consistently, and 94% of individuals said they use AI daily or often. Not one respondent said they don't use it at all. 2 Productboard also reports that 98% of respondents have changed, or plan to change, team structures because of AI, and its findings include a claim of more than 33 hours saved. 2

Productside's annual State of AI for Product Management report, based on more than 250 product professionals, gives a lower figure: roughly four in five PMs use AI every day. 3 Its headline finding is about what happens beyond the individual. Only about one in four respondents work somewhere with an actual AI strategy behind that usage. 34 The report puts it plainly: your team is using AI, but your organization isn't ready for it. 3

Where the numbers diverge

The gap between 80% and 94% probably reflects different samples and question wording more than a real contradiction. Productboard's respondents are enterprise-focused, while Productside draws from a wider range of industries, regions, and company sizes. 23 There is also some confusion about attribution. One analysis credits the 96% "frequent usage" figure and the "deeply embedded" language to ProductPlan research. 4 Productboard publishes nearly identical numbers under its own name. 2 Readers should treat these figures as directional rather than precise.

The pattern is consistent across all of them, though. Individual adoption is effectively saturated. Organizational maturity lags far behind. One synthesis of the Productboard data puts the share of teams treating AI as a "core, strategic capability" at only around 6%. 1

The ROI question

The strategy gap matters because of a broader debate about returns. Widely cited MIT research found that roughly 95% of enterprise AI pilots produced no measurable ROI. Deloitte's 2026 enterprise study, by contrast, found 66% of organizations reporting tangible gains. 4 These findings can coexist. Real value is being created in some places, while a great deal of activity is going into pilots that never connect to business outcomes. 4

For product teams specifically, that disconnect is sharper because expectations are rising. Airtable's predictions report says 76% of product leaders expect their AI investment to grow next year. It also says 92% of product leaders now own revenue outcomes, more than double the share from a few years ago. 5 More spending and more revenue accountability, with only a quarter of organizations working from a coherent strategy, is a recipe for the stalled-pilot pattern MIT describes.

Tool sprawl adds to the risk. Productboard found that 88% of product teams use two or more AI models, and 60% use two different tools for AI prototyping alone. It frames this fragmentation as a source of new risk. 2 When teams pick tools independently and nobody owns policy, measuring impact becomes hard. That is exactly the ownership and measurement breakdown Productside set out to examine. 3

What AI is actually doing

The most useful framing may be about task type. According to the DEV Community analysis, AI has absorbed the coordination work that already consumed around 60% of a PM's week. That includes drafting PRDs, building presentations, competitive research, and synthesizing feedback. 1 Airtable's list of use cases is similar: writing briefs in minutes, mapping roadmap items to goals, automating launch updates, and managing budgets. 5

What AI has not absorbed is judgment about what is worth building, and the work of defending that call with evidence. 1 The same analysis points out that 43% of startups fail because they built something nobody needed, not because they built too slowly. 1 Speed gains on documentation don't address that failure mode.

The reading

The evidence points to a clear conclusion. The adoption story is over, and the strategy story has barely begun. AI has made product managers faster at the parts of the job that were never the bottleneck. Most organizations have not yet translated that speed into better decisions, standardized tooling, or measurable returns.

Airtable predicts that AI-powered product strategy will become the standard. 5 If so, it will be because companies close the ownership and measurement gaps Productside identifies, not because more people open a chatbot each morning. For now, the competitive edge is likely to belong to the minority of teams that treat AI as an organizational capability rather than a personal productivity habit.