TL;DR
The short version
Most companies still treat AI like a tool-selection problem: pick a vendor, buy seats, wait for value. KPMG's Q2 2026 survey says that's the wrong altitude. The clearest predictor of getting a return wasn't the model or the budget — it was accountability.
Organizations with a clear owner for AI decisions were about 3x more likely to report ROI, and the effect was far stronger when the CEO held the outcome. The figure is KPMG's own survey data — read the 3x as a strong correlation, not a proven law — but the pattern underneath it is hard to argue with.
Drawn from The AI Daily Brief (Nathaniel Whittemore) unpacking the KPMG Q2 2026 AI Quarterly Pulse Survey.
Most companies are solving the wrong problem
When AI is an IT line item, no one above the team that bought it feels responsible for whether it pays off. The pilot runs, the demo impresses, and then it quietly stalls — because nobody owns the gap between "it works" and "it changed the business."
On The AI Daily Brief, Nathaniel Whittemore frames the survey's lesson through Ethan Mollick, who names the shift directly:
Decisions about how to use AI in your organization are increasingly organizational design and strategy decisions, not IT choices. How do you integrate agents into your firm? What intelligence will you outsource? What are the boundaries of the firm? What is the role of people?
Ethan Mollick
Those aren't procurement questions. They're CEO questions. And the survey shows what happens when a CEO actually answers them.
What CEO ownership does to the numbers
The split is stark. Where the CEO is accountable for AI, 14% of organizations report established ROI; where they're not, it drops to 4%. Ask whether AI is delivering meaningful business value and it's 57% versus 21%. Ask about confidence in future-proofing the AI strategy and it's 60% versus 22% — all per KPMG's Q2 2026 survey.
Same technology. Same market. The variable that moves is who's holding the outcome.
This isn't about the CEO writing prompts. It's about a single, visible owner who decides what gets integrated, what gets cut, and what "good" looks like — instead of letting a dozen teams each run their own unaccountable experiment. KPMG found few companies have a single point of accountability; the finding isn't that you must centralize it, but that clarity beats fog.
The boring wins are now table stakes
The survey also caught a quieter shift: priorities moving from efficiency to opportunity. Productivity gains as a top priority fell from 42% to 35%, cost reduction slipped from 31% to 29%, and faster decisions dropped from 41% to 36%.
That doesn't mean efficiency stopped mattering. It means leaders stopped treating it as the prize. The companies pulling ahead are aiming AI at new value, not just cheaper versions of old work.
Discipline is replacing FOMO, too. About half of organizations re-phased an AI deployment when costs outweighed value — not retreat, but a sign teams are measuring instead of believing. The weak spot: only about a third have full visibility into their AI operating costs, which gets dangerous as usage-based pricing replaces the subsidy era.
Make accountability explicit — then do the harder work
If you want one quick win from this, it's cheap to start: pick the AI decisions that matter — what to integrate, what to outsource, where humans stay in the loop — and assign a clear owner to each.
One caveat worth holding onto: leaders consistently overestimate how the rest of the company feels. The same survey shows the employee side moving the other way.
KPMG Q2 2026 AI Quarterly Pulse Survey (figures confirmed in the published report)
Ownership at the top sets direction. It doesn't manufacture buy-in at the bottom. The takeaway is simple enough to act on this week: AI value is an org-design decision, not a software purchase — and if no one owns the outcome, you've likely already capped your return.
