AI Use Hits 91%, but Fewer Marketers Can Prove It Pays

Adoption is no longer the question. By one large survey, nearly nine in ten marketers use AI as part of their daily work, and most say it gets work out the door faster. The harder question has moved to the next meeting, where someone asks what the spend actually returned.

On that question the data is moving the wrong way. More teams are using the tools, and fewer of them can show the result.

Usage Up, Proof Down

Jasper's State of AI in Marketing 2026, a survey of 1,400 marketers, reports that 91% actively use AI, up from 63% the year before. It also reports that only 41% can prove AI return on investment, down from 49% last year. Half say they are bringing work to market faster, 45% have lowered operating costs and 65% of teams now have designated AI roles.

Jasper reads the dip as a sign of raised expectations: leadership now wants business outcomes, not hours saved. That fits the pattern in Basis's roundup of AI ROI research, which cites IAB data showing only 40% of marketing professionals use or plan to use defined KPIs for AI tools. A team without a KPI cannot lose an argument about ROI, but it cannot win one either.

The Bottleneck Is Review, Not Software

Jasper names the top three obstacles to scaling AI as brand, legal and compliance reviews, lack of output quality, and data and privacy risks. None of those is a model problem. They are process problems: who checks the output, against what standard, and how long that check takes.

That matters for measurement because a review queue is where time savings quietly disappear. If a draft takes ten minutes to generate and four days to clear legal, the productivity figure in the dashboard and the speed the campaign actually shipped at are two different numbers. Teams that only count the first will report a gain that nobody else in the business can see.

Data Ownership Sits Outside the Team

The Supermetrics 2026 Marketing Data Report, which surveyed 435 marketing professionals between October and December 2025, finds that only 6% have fully embedded AI into their workflows even though 80% feel pressure to adopt it. It also finds that 52% say data strategy and measurement decisions are made by external teams, and that only 31% confirm CMO involvement in those discussions.

The same report puts the cost in plain terms. 45% name measuring ROI as their number one challenge, 40% struggle to prove it across channels, and only 33% say they can activate their data effectively. The most common measurement method is still A/B testing, used by 69% of respondents, which is a reminder that a controlled comparison remains the most trusted proof the discipline has available.

A Playbook for Proving the Return

• Set the KPI before the pilot. Write down the metric, the baseline and the target for each AI use case before it launches, so a result can be judged against something agreed in advance.
• Measure cycle time end to end. Track the whole path from brief to live, including review and approval, not just the minutes saved at the drafting step.
• Hold out a control group. Run AI-assisted and conventional versions of the same creative or journey side by side, and compare conversion and revenue rather than output.
• Fix the review path. Agree brand, legal and compliance criteria up front, and give reviewers a checklist so the queue shortens instead of growing.
• Put marketing in the data room. If measurement decisions sit with external teams, secure a seat so the definition of success is not set without you.
• Report returns in business terms. Lead with revenue, retention and cost per outcome, and keep hours saved as supporting evidence.

The 60% of measuring marketers who report returns of 2 to 3 times or higher suggest the technology is paying off for those who check. The marketers falling behind on the 41% are mostly not failing at AI. They are failing to show their work.