90% of CEOs Say AI Has Had Zero Effect on Their Business — But That Number Tells a Much Darker Story

I just came across a survey result that should be front-page news but somehow isn't: the National Bureau of Economic Research surveyed executives and found that more than 90% reported no effect on their firm's employment from three years of AI. And 89% said they saw no effect on productivity either. After billions spent on AI tools and subscriptions, nearly every C-suite executive is essentially saying: nothing changed. I need to talk about why that number is both reassuring and terrifying at the same time.

Wait — That Can't Be Right

I know what you're thinking. Three years of ChatGPT dominating every business meeting. AI coding assistants, AI writing tools, AI customer service bots. Entire job categories supposedly being disrupted overnight. And 90% of executives say... nothing? Either the executives are fibbing, the survey methodology is broken, or something deeply strange is happening in corporate America. Honestly, probably a mix of all three.

Let me break down why this number is actually more alarming than reassuring — and what's really going on beneath the surface.

The Productivity Paradox, Reloaded

Economists have a name for this phenomenon: the productivity paradox. We've seen it before, and it's the most important context for interpreting the NBER results. In the 1970s and 1980s, companies invested massively in computers. For over a decade, economists found almost no measurable productivity improvement. Then in the 1990s, it exploded — suddenly and dramatically. The gains were real, but delayed by the time required to fundamentally redesign business processes around the new technology.

We may be living through the same pattern with AI. The tools are deployed. The subscriptions are paid. But actually changing how a law firm does discovery, or how a hospital system processes records, or how a manufacturer designs supply chains — that takes years of process redesign, training, and organizational change management. We're not there yet. Most companies are still in the "bolt AI onto existing workflows" phase rather than the "rebuild the entire process around AI" phase.

Three Reasons the Survey Results Make Sense

First: AI is a long game. Most AI deployments over the last three years have been in the "assistant" phase — helping individual workers do discrete tasks faster. A lawyer who uses AI to draft motions 3x faster is more productive. But if the firm doesn't restructure how many lawyers it hires or how many billable hours it charges, that productivity gain is invisible to traditional metrics. The transformational effect requires process redesign, which takes years.

Second: Measurement is fundamentally broken. GDP, productivity statistics, and employment figures were designed for a manufacturing economy. When knowledge work gets faster, it often doesn't show up in the numbers. A software engineer who writes twice as many features in the same number of hours hasn't changed their "output" in any way our current statistics capture. The gains are real but unmeasured.

Third: Most AI implementations are still shallow. Buying GitHub Copilot or a ChatGPT Enterprise license is not the same as AI transformation. The organizations seeing real gains aren't the ones who subscribed to AI tools — they're the ones who fundamentally redesigned workflows around AI capabilities. That kind of organizational change requires executive commitment, change management, and time. Most companies haven't done it yet.

The Employment Number That Should Scare You

Here's the part of the "no effect on employment" finding that I think is being widely misunderstood. Aggregate numbers hide the most important dynamic. If a company would have hired 200 people for a function in 2026 but only hired 120 because AI picked up the productivity slack — that's a massive employment effect that doesn't show up as "job losses" in any traditional survey. The executive will truthfully report: we didn't lay anyone off. What they won't volunteer: we also didn't hire the 80 people we would have needed before AI.

The jobs most at risk from AI aren't necessarily the jobs being eliminated today. They're the jobs that won't be created tomorrow. That's a much harder thing to measure and a much easier thing to ignore.

What Comes Next

The productivity paradox historically resolves — suddenly and dramatically. When it does, the companies that figured out AI implementation first will have a significant, durable competitive advantage. The survey results showing "no effect" are not evidence that AI doesn't work. They're evidence that we're still in the installation phase.

If you're a worker in a field that AI can affect: the three-year window of "nothing's changed" is ending. The organizations that are figuring this out are starting to pull ahead. Don't mistake "it hasn't happened yet" for "it won't happen." History says otherwise.

If you're an executive: you're in the majority not seeing ROI. But the companies that crack this first won't share their playbook. Urgency matters here more than most business leaders currently appreciate.

What's your experience? Drop a comment below! 👇 Has AI actually changed how your company operates day-to-day, or does the "no effect" finding match your reality? I want to hear from people in the trenches — what are you actually seeing?