56% of CEOs Say AI Isn't Making Them Money. I'm Not Surprised.

PwC just published their 2026 CEO Survey, and one number is making the rounds in every boardroom conversation I'm having right now: 56% of CEOs report zero increase in revenue or decrease in costs from AI over the past year. Zero. After billions spent on licensing, deployment, change management initiatives, and "Prompting 101" lunch-and-learns.​

Here's what's not making the rounds: the reason why.

The Part of the Story Nobody's Telling

Every time a stat like this drops, the hot take machine kicks into gear. AI is overhyped. The bubble is bursting. The technology isn't ready for enterprise.

That's the wrong diagnosis — and leaders who buy it are going to make the wrong call.

Deloitte surveyed over 3,200 business and technology leaders across 24 countries this year. 74% said they want AI to grow revenue. Only 20% have actually seen it happen. A 54-point gap between ambition and reality — and it has almost nothing to do with the quality of the tools.​

DataCamp's 2026 research is even more direct: organizations that treat workforce capability as core infrastructure — not an afterthought — are significantly more likely to see meaningful AI ROI. The differentiator isn't the software. It's what happens after the software gets deployed.​

The CEOs reporting zero ROI didn't buy bad tools. They solved the wrong problem.

What's Actually Happening Inside These Organizations

There are two failure modes I see consistently in enterprise AI rollouts, and both of them produce flat ROI — but for completely different reasons.

The first is the Shelfware Stall. Low utilization, low value. The tool got deployed, the announcement email went out, and then... nothing. People went back to doing their jobs the way they always have. The licenses sit idle. The investment collects dust.

The second is trickier — and honestly, more dangerous — because it looks like success from the outside. I call it The Activity Trap. Utilization is high. People are logging in, prompting away, generating summaries and drafting emails at impressive speed. The dashboard looks great.

But not a single core KPI has moved.

Microsoft's own data tells this story. Copilot has 15 million paid enterprise seats and Satya Nadella called it "a true daily habit" on the Q2 earnings call. In the same report, analysts noted that the average user estimates AI saves them about 20% of typical task time — but most cannot quantify what that translates to in revenue or cost savings.

That gap — between time saved and value created — is exactly where the 56% live.

Here's My Stance

Saving time is not a business outcome. It's a precondition for one.

If your team is using AI to do the same work faster, but that freed-up capacity isn't flowing into something that moves a KPI — a sales goal, a retention rate, a margin target — then you haven't transformed anything. You've just added a faster gear to the same machine.

The organizations in that top 12% — the ones PwC found that saw both revenue growth and cost reduction from AI — didn't get there by deploying better tools or running more training sessions. They got there by answering a harder question before the rollout ever started: What does winning actually look like for each team, and how does AI get them there?​

That question changes everything — what you train, how you measure it, and whether your CFO is still funding it in Q4.

So Before You Forward the PwC Report to Your Leadership Team...

Ask yourself which number you're actually in. Because "we have AI tools and people are using them" is not the same as "we are in the 20%."

I built a free AI Adoption Audit for exactly this moment — when the data is uncomfortable and you're not sure whether your organization has a utilization problem, a value problem, or both.

Take the free AI Adoption Audit →

Four questions. Three minutes. You'll get a specific diagnosis — not a generic score — and a tailored roadmap based on exactly where your rollout stands today.

The 56% isn't a technology story. It's a strategy story. And strategy is fixable.