Your CEO Will Ask What AI Is Delivering. Have These Four Answers Ready.

If you own AI automation at your company, you're about to be asked what it's delivering and how it's being measured. McKinsey's July 2026 analysis found that 62% of organizations have moved beyond experimentation to active deployment, and 93% report exceeding their AI budgets. Once AI is in production and over budget, the questions come from the top. Here are the four I'd prepare for.

"What are we spending, and is it on budget?"

You aren't alone if the honest answer is "more than planned." A WitnessAI survey of 300 executives, reported by CFO Dive this July, found that 68% said at least some AI initiatives ran over budget in the past year, and a third said overruns happened mostly or always. A good answer has four components: every AI tool, who owns it, what it costs, and which workflow it serves. If assembling that answer takes more than a day, that's a finding in itself.

"What did we get for it?"

This is the hard one. The same CFO Dive report cites a CloudZero survey in which 87% of finance leaders felt pressure to connect AI spending to business outcomes within the next year, and only 22% had done so.

The way to get ahead of it is to pick, for each workflow, one measure that finance already trusts, and record it before launch. For follow-up, that might be the time from a call to a sent email. For renewal prep, it's the renewal rate on the accounts the workflow covers. Without a starting number, you can't show movement, and a story merely about time saved likely won't survive the budget discussion.

"What are we stopping?"

Emergn's research, covered by CIO Dive, found that U.S. organizations lose an average of 2.4% of annual revenue on AI initiatives that fail to deliver expected value. Only 30% said shutting down an underperforming AI initiative is normal practice, and nearly half said projects typically stop only after significant time and money have been spent.

The fix is to decide the stopping rule before you start. Write down what the measure should do by day 60 or day 90, and agree in advance that you'll stop or change course if it hasn't. A project with an exit rule is easier to approve, because it has a stop-loss attached to it.

"Where could we spend less without losing results?"

McKinsey's analysis suggests the opportunity is real. It says organizations that build the habits to optimize spend, improve accountability and redirect savings toward the highest-value uses could reduce their AI costs by 20% to 30%. In a revenue team, the easiest places to look are overlapping tools and workflows that never moved their measure. Moving that money to the one or two workflows that did is a stronger pitch than asking for new budget.

Walk in prepared

If you can answer those four, you've turned AI from a line item into a set of outcomes you manage. The review becomes a decision about where to put the next dollar. If you'd like help setting measures for your first workflows, that's a good use of an hour with us at Rampwell.