A laptop showing revenue and analytics charts on a desk

Dedicated or bolted on? What new Wharton research says about the Chief AI Officer

Most large companies now say their AI spending is paying off. Far fewer say it is paying off well. That gap is the heart of this year's AI adoption study from Wharton Human-AI Research and GBK Collective, and one of the clearest signals in it concerns the Chief AI Officer.

The report, The Adaptive Advantage: Who's Winning With AI and How, was published on 13 October 2026. It is the fourth year of the study, and this time it surveyed 1,007 US enterprise leaders, from managers up to the C-suite. Here is what it says, and what it means for anyone running AI at the top of an organisation.

Positive returns are common. Strong returns are not.

According to the study, 80% of leaders report positive returns on their AI investments, but only 28% describe those returns as strongly positive. Expectations are still high: 83% expect AI to fundamentally transform or majorly affect their industry within two to five years, and 87% expect their AI budgets to rise over the next 12 months.

So the question for a Chief AI Officer is no longer whether AI works. It is why it works so much better in some places than in others.

A dedicated seat makes a difference

This is the finding that should get the attention of every board. At organisations with a newly designated Chief AI Officer, 58% of leaders report strongly positive ROI. Where AI oversight has been added to an existing executive's job, that falls to 24%. With no such role and none planned, it drops to 8%.

The report also notes that two thirds of organisations now have a CAIO or equivalent, but nearly all of this year's growth came from handing the job to existing executives rather than creating a dedicated role. In other words, plenty of companies have ticked the box without giving anyone the time and authority to do the work.

The authors are careful to say these are self-reported, correlational findings. A dedicated CAIO will not deliver returns on their own. But the pattern is strong, and it points the same way as everything else in the report: clear ownership comes before strong results.

Pick an approach and stick to it

Platform strategy follows a similar shape. Half of organisations that have standardised on a single approved AI platform report strongly positive ROI. That compares with 30% of those that allow several approved tools under a formal policy, and 9% of those still experimenting with no defined approach. The report's reading is that any settled position beats none.

For a CAIO, this is a practical first job. Decide how tools get approved, write it down, and make sure people know.

Measure growth, not just hours saved

Hours saved is the most common way departments measure AI, used by 40% of them. Net-new revenue is the least used of the major measures, at 27%.

That choice seems to matter. Among departments that track at least one growth measure, such as new revenue, sales conversion, output or customer satisfaction, 35% report strongly positive returns. Among those measuring efficiency alone, it is 13%.

As Wharton professor Sonny Tambe puts it in the report: "Companies believe AI is going to transform their industries, but many are still asking it to prove itself by saving time. The companies seeing stronger returns are going further. They are redesigning workflows, putting clear ownership and governance in place, and measuring AI against outcomes that matter to growth."

The supervised middle is where returns are weakest

Agents are now common, but few are trusted to act. Brian Smith of GBK Collective notes that two thirds of departments say they use AI agents, yet only 17% let an agent complete a task without a person approving each step. "Most companies are sitting in a supervised middle, paying for autonomy and then reviewing every action, and that is where returns are weakest," he says.

The report suggests giving agents more meaningful work and using automated monitoring rather than step-by-step human sign-off. Departments that monitor agents automatically have transformed their workflows at 24%, against 9% of those that do not. That is not a case for removing oversight. It is a case for designing oversight that scales, which is exactly the kind of decision that needs a named owner.

Five things a CAIO can take from this

  1. Make the role real. If AI sits on top of someone's existing job, be honest about whether they have the time and mandate to run it. The gap between 58% and 24% is hard to ignore.
  2. Settle the platform question. A single platform or a formal multi-tool policy both beat open-ended experimenting.
  3. Add a growth measure. Keep tracking time saved, but add at least one measure tied to revenue, conversion or customer satisfaction.
  4. Move pilots into production. The report recommends fully embedding AI in core workflows and moving more than half of pilots into production, rather than bolting AI onto a few processes.
  5. Rethink how agents are supervised. Approving every step is expensive. Automated monitoring, with clear limits on what agents can do, is where the report sees stronger results.

The bottom line

The Chief AI Officer role is no longer a novelty. What this research suggests is that the title alone is not enough. The organisations getting strong returns have given someone real ownership, settled on an approach, and started measuring AI against the outcomes the business cares about. For CAIOs heading into 2027 planning, that is a useful checklist.

Sources

Image: Antoni Shkraba on Pexels (Pexels License).