For CEOs

What a CEO needs to know when AI spend isn't reaching the P&L

For most companies the tools are in use and the return has not arrived, because spending and usage are the only two things the reporting counts. The return is decided after usage, by three conditions a standard report does not measure: Direction, Skill and Reinvestment. Finding which one is failing is the CEO's first job, and it starts with people.

Take the free AI Profit Readiness Assessment

The question, in the CEO's words

“We've spent the money and people are using the tools. I can't point to a number that moved. Is this working, and how would I know?”

Why the dashboard looks healthy

Your reporting counts what you spent and how much the tools are used. Both are easy to count, and in most companies both look fine. Usage is rarely what stands between you and the return any more. BCG's 2026 AI at Work survey of 11,749 workers found that 74% of frontline employees are now regular AI users, up 23 points in a year (BCG, June 2026).

So when the board asks what AI has done for the business, the honest answer is often that people are busy with it and nobody can say what it changed.

Where the return is decided

The AI Profitability Gap™ is the distance between what a company expected AI to deliver and what it is delivering. Underneath it is a chain of five links: Investment, Usage, Direction, Skill and Reinvestment. Together they produce Return. Your reports cover the first two links. The other three make up the AI Profitability Gap.

Direction. Is the AI work pointed at an outcome the business already measures?
Skill. Can your people get more out of the tools than the first obvious use?
Reinvestment. When AI saves an hour, has anyone decided what that hour is now for?

BCG found that a clear AI strategy lifted measurable business impact by 25 percentage points, against about five points for better tools alone (BCG, June 2026). That is a Direction finding, and Direction starts in your office.

Why this lands on your desk

Each condition sits with a different seat. Direction is a leadership decision. Skill usually sits with the Chief People Officer. Reinvestment sits with whoever owns the work AI is landing on, which is often the COO or the CMO. When nobody holds the whole chain, each executive reports their own piece as healthy and the return still does not show up.

The CEO is the one seat that sees all five links at once, which is why the question keeps coming back to you.

What to do first

1. Ask each member of your leadership team which business number their AI work is meant to move this quarter. If the answers differ, start with Direction.
2. Ask where the hours AI saved last quarter went. If nobody knows, Reinvestment is open.
3. Take the free AI Profit Readiness Assessment, and ask your leadership team to take it too. It is eight questions, takes about two minutes, and shows which condition is costing you the most.
4. If the read points to your people and process, bring the team into the AI Profit Workshop: a founder-led half-day for up to six people, $2,999, credited in full toward the AI Profit Sprint if you start one within six months.

Questions CEOs ask

How do I know whether AI is paying off in my company?

Look for a business measure that has moved since the AI work started, against a baseline taken before it. If spending is approved, usage looks healthy and nobody can name that measure, the return is being lost somewhere between usage and the result.

Who on the leadership team should own the return on AI?

The CEO owns the whole chain, because the three conditions sit with different seats. The Chief People Officer usually owns Skill, the COO or CMO owns Reinvestment in their function, and IT leadership owns the platform alongside them.

Is the problem our AI tools?

Usually not. Most companies already have capable tools in daily use. Average Robot works on the people and the process around the technology you have bought, and where the technology itself has to change, we say so and bring in a build partner.

What should I tell the board while this is being fixed?

Show the chain: what was invested, how widely the tools are used, and which of the three conditions you are now measuring. Our guide to reporting AI ROI to the board sets out a one-page structure.

Next step

Eight questions and about two minutes, and you get a read on which of the three conditions is costing you most, with a first move.