McKinsey's 2026 AI survey: what it means for a business your size

McKinsey's yearly AI survey came out at the end of August, and one line from it keeps coming back to me. Plenty of firms use AI now. Far fewer make money from it.

The survey asked 1,719 people in 97 countries. More than a third work for companies turning over $1 billion or more, so this is mostly a big-company picture. The pattern still applies to a clinic, a dental practice or a building firm.

The gap is widening

Big companies are moving faster with AI agents, the tools that carry out a few steps on their own rather than just answering a question.

"Forty percent of respondents from large organizations ... report scaling AI agents, up from 27 percent last year."

For smaller organisations the figure "remained flat at 22 percent". So the big firms grew and everyone else stood still.

Using AI is not the same as profiting from it

Only 37% of the people asked said AI had any effect at all on their profit. The group McKinsey calls "AI high performers", where AI drives at least 5% of profit, is 6% of respondents, "unchanged from 2025".

So nearly everyone has tried it, about a third see some return, and a handful see a real one.

What the 6% do differently

This is the part worth your time. Nearly three-quarters of the high performers say they have fundamentally redesigned workflows because of AI. Among everyone else it's about a quarter.

In plain terms, the winners didn't add a chatbot to the way they already worked. They changed the way the work flows: who gets the enquiry, what gets asked, what happens next and who signs it off.

What I'd take from it if I ran a practice or a trade business

Pick one process that eats time every week. New enquiries are usually the obvious one: someone asks a question at 9pm, nobody replies until the next afternoon, and by then they've booked somewhere else.

Write down how that enquiry moves today, step by step. Then decide which steps a tool could take, which need a person, and what the person should see when it lands on their desk. Only then pick the tool.

Doing it in that order is what separates the 6% from everyone else. It's also why we start every automation project with one task and the process behind it, before any software is chosen.

One more figure from the survey: 32% of respondents said their organisation decided not to buy at least one software product because they could build it themselves with AI coding tools. That's mostly large teams with developers on staff. For a business without one, the lesson is the same either way: the tool matters less than knowing exactly what you want it to do.

If you'd like to see what that looks like for your own enquiries, message us and we'll send you an example of how it would work for your business.