Two-thirds of IT leaders see AI results. Only eight would call the CEO.

By Sophie Lindqvist

Tech entrepreneur Azeem Azhar polled 160 IT vice presidents in a Las Vegas conference room and drew two very different hand counts. Asked who had measurable AI results, two-thirds of the room raised a hand. Asked who had a win strong enough to interrupt the CEO's summer vacation, only eight did.

The exercise isn't peer-reviewed. It's also hard to argue with. It puts a clean, public number on the gap between AI adoption and AI urgency inside the enterprise — and that gap is where the AI bubble debate now lives.

What did Azhar actually ask?

Azhar's drill was simple. First question, soft: who has measurable AI results to show? About 107 of the 160 leaders raised a hand. Second question, hard: who has results good enough to interrupt the CEO's summer vacation? Eight hands went up.

The two questions are a stress test. The first measures what IT can point to. The second measures what IT would feel comfortable calling urgent. That gap — between "we have a story" and "we have an emergency" — is where most enterprise AI still sits.

Most senior IT leaders can name at least one AI deployment that moved a number. Far fewer feel the case is boss-worthy.

Why does the gap matter?

That mismatch is the structural question behind the AI bubble debate. Billions are flowing into AI applications, infrastructure, and talent. The question — the only question, really — is whether the in-flight results justify the spend.

The original write-up of the poll framed it this way: "Whether that slow pace of progress justifies the massive investments is one of the key questions in the AI bubble debate." Azhar's room of 160 is anecdotal, but the pattern rhymes with what broader enterprise surveys keep showing. AI is landing somewhere. AI is not yet landing urgently.

If measurable AI wins are common and boss-worthy ones are rare, the value being captured sits lower in the org chart than the spending suggests. That's the worry for anyone funding AI at scale — and the reason a CEO might, fairly, want a stronger answer than "two out of three hands went up."

What's the spending backdrop?

The Azhar anecdote reads oddly against the actual money flowing in. A separate analysis from the financial management platform Ramp — built on spending data from tens of thousands of companies and billions of transactions — shows companies have significantly increased their outlay on AI applications.

The macro picture and the room-level picture diverge:

• Macro: AI budgets are growing across the enterprise.
• Room level: The urgency inside IT to escalate AI wins is not.

CEOs are approving spend at a pace outpacing the wins their own IT vice presidents would feel comfortable flagging. That's not a critique of the technology. It's a critique of the timing. The spending is running ahead of the case. Either the case catches up, or the spending slows down.

What counts as "measurable"?

Azhar didn't define measurable for the room. In practice, the term covers a wide range of things an IT shop might already be doing:

• A chatbot that deflects tier-one support tickets.
• A forecasting model that tightened inventory by a few points.
• A back-office workflow that now runs without a human in the loop.
• A coding assistant that shaved time off routine pull requests.

Most IT shops can point to at least one of these. Rare, per the poll, is a result that is both measurable and decision-changing at the executive level. The kind of win that reorders priorities. The kind worth pulling a vacationing CEO back for.

That's where the eight hands fell. It's also where the line gets drawn between AI as an incremental tool and AI as a strategic bet. Plenty of leaders are in the first camp. Far fewer are in the second.

So what happens next?

The poll captures enterprise AI in a strange middle state. The technology is past the proof-of-concept phase for most IT shops. It is not yet at the strategic-emergency phase for most CEOs. That gap is exactly where the AI bubble question will be litigated over the next two quarterly earnings cycles.

If Azhar reruns the drill six months from now and the boss-worthy hands grow — even modestly — the spending justifies itself. If the number stays parked at eight of 160, the bull case gets harder to defend. Either way, the result gives the industry a clean, immediate measure of where enterprise AI actually stands — and a sharper yardstick than another vendor deck.