Wharton Human-AI Research and GBK Collective released the fourth edition of their enterprise AI adoption study on October 14. Among managers and above whose company has a newly designated chief AI officer, 58% report strongly positive AI returns. Where the duties were added to an existing executive, the figure is 24%. Where no chief AI officer exists and none is planned, it is 8%.
What the study measured
According to the report, “The Adaptive Advantage: Who’s Winning With AI and How,” the researchers ran a 15-minute online survey of about 1,400 full-time employees at U.S. private-sector companies with more than 1,000 employees and more than $50 million in revenue. Interviews took place between July 9 and July 29, 2026. For the first time the sample includes entry- and mid-level staff (n=409) alongside the managers and above (n=1,007) who answer the ROI questions.
The headline return numbers are modest. The report says 80% of managers and above now report positive returns from AI, and 28% report strongly positive returns. It also says 93% formally measure ROI at least occasionally, so most of these companies are measuring and still finding strong returns scarce.
Three findings a buying committee can act on
The first is the chief AI officer split. The report says 67% of managers and above have a chief AI officer or equivalent, up from 61% in 2025 and 47% in 2024. But 42% describe the duties as added to an existing executive’s role, up from 34% a year earlier, while 25% describe a newly designated position, against 26% in 2025. The researchers write that “nearly all of this year’s growth is reassignment rather than dedicated appointment.” Another 24% say they plan to add the role, down from 31%.
The returns follow the structure of the role. In the report’s words, “A shared title does not.” Companies that gave an existing executive the title report 24% strongly positive returns, which the report describes as no better than organizations still deciding. The 8% figure for companies with no role and no plans rests on 63 respondents, and the authors flag it as a small base to be read directionally. The 58% and 24% figures rest on 252 and 421 respondents.
The second finding concerns platforms. Only 19% of managers and above who influence AI decisions say their organization standardizes on one approved AI platform or vendor. Another 49% allow multiple approved tools under a formal policy, 20% run a central hub or orchestration layer, and 9% are still experimenting. Strongly positive returns by approach:
- One approved platform: 50% (n=184)
- Multiple approved tools under a formal policy: 30%
- Central hub or orchestration layer: 19%
- Still experimenting, not standardized: 9%
The report summarizes it this way: “Standardizing to one approved AI platform is optimal, but the main goal is any settled approach.” A central hub, the architecture many vendors currently sell as the answer to tool sprawl, sits below formal multi-tool policies in this sample. The report does not explain why, and the sample cannot say whether hubs underperform or whether companies adopt them after returns have stalled.
The third finding is spend visibility. Of managers and above who are very confident their organization can track total AI spend, 56% report strongly positive returns. Among those less confident, 10% do. The authors caution that these respondents are also more enthusiastic about AI in general, so some of the gap may reflect conviction rather than capability. The report also lists the cost of tools, compute and licenses as the leading barrier to stronger returns, at 34%, and says 87% expect budgets to rise over the next 12 months.
How this sits next to other counts of chief AI officers
Surveys of chief AI officer prevalence disagree widely, and the gap comes from who is asked. IBM’s Institute for Business Value reported on May 4 that 76% of the organizations in its study of 2,000 CEOs have a chief AI officer, up from 26% a year earlier. Wharton and GBK count 67% among U.S. enterprise managers and above, and only 25% of those describe a newly designated position. IBM asked CEOs worldwide and Wharton asked U.S. enterprise staff, so the two numbers cannot be compared directly. Both studies report fast growth in the title.
Two earlier pieces on this site bear on the finding. Our look at Target’s first chief AI officer and the retail pattern covered a dedicated appointment, and Deloitte’s agentic AI survey found 42% of companies had tested agents while only 15% had scaled them. Wharton’s data points to two structural variables worth checking against that gap: who owns the work and how many platforms the company runs.
What it means for the buying committee
For a CIO or CAIO, the practical reading has three parts. A survey of this kind shows association and the authors say as much, so none of these steps is a promise of returns.
- Check whether the chief AI officer duties belong to someone whose job is only AI. If the title sits on top of a full CIO or CTO workload, the study places that company with the 24%.
- Count the approved AI platforms and name the exit date for each tool that is not on the standard list. In the survey, 49% of respondents allow several tools under policy and only 9% of the still-experimenting group report strong returns.
- Ask finance whether total AI spend, including tokens, licenses and API usage, can be reported in one number. The 56% versus 10% split is the largest gap in this part of the report, even after the authors’ caveat.
Vendors pitching a platform, a hub or a new executive seat to a committee can be asked to show returns by the same cuts the study uses: dedicated versus shared ownership, one platform versus many, and tracked versus untracked spend. A committee that arrives at a renewal with its own answers on those three points is in a stronger negotiating position than one that relies on a vendor’s case studies.
