Let me ask you the question I’ve been putting to owners all year. When AI takes over part of somebody’s job, what’s your first instinct? Cut the role, or teach the person in it? Most owners won’t admit to the first answer. You can still see it in the math they’re running in their heads, and I understand why. Payroll is the biggest line on most small business P&Ls, and every week another headline promises AI will shrink it.
So let’s look at what businesses using AI are doing, as opposed to what the headlines predict. The Federal Reserve Bank of New York (opens in new tab) has surveyed firms in its region about AI for three years running. In its August 2026 survey, 61% of service firms said they now use AI, up from 40% a year earlier. Only 4% of those service firms reported laying anyone off because of it. Just over a third retrained their people instead, and the researchers concluded that retraining is still the main way businesses are adjusting their workforces to AI.
That lines up with what I tell my clients. I don’t think AI makes people less important. If you learn how to use the tool, it can make people even more important. The owners who treat AI as a way to get rid of people are usually trying to fix the wrong thing, and I watched that same mistake cost businesses long before anyone had heard of ChatGPT.
I don’t think that AI makes people less important. I think if you learn how to use the tool, it can make people even more important.
— Gary Henson, Founder of Gary Henson Group
Earlier this year I took on a client with about 100 employees. Leadership told me they wanted a better culture, one where people understood their roles and would be more productive because of it. Then I asked the question I ask every client: “What do you want?” It turned out culture wasn’t the goal. They wanted higher profits, more output per person, and a map of where the gaps were between departments. What was missing was the president’s commitment to having happy employees, people who looked forward to coming to work and told their friends and family it was a great place to work. It wasn’t. Their people were productive in spite of the environment around them. I told them plainly that I didn’t know if I could help them as much as they wanted.
I think about that client every time an owner asks me whether AI means they can run leaner. An owner who sees employees as a cost to squeeze will see AI as a better way to squeeze. And the people left behind will notice. Your employees are the internal customer of your business. When they watch a coworker get replaced by software without anyone offering to teach that coworker the software first, they draw one conclusion: they’re next. Then they start acting like it.
Here’s my view on where AI belongs. Using it to be more efficient in the most literal sense, having it answer your phones or do the work for you, may not be worth much. Using it as a tool to grow your business and develop your culture is a much stronger way to leverage it over the long run.
There’s a second reason to be careful, and it comes out of Stanford. The Stanford Digital Economy Lab’s August 2026 update to its “Canaries in the Coal Mine” research (opens in new tab) found that employment for workers ages 22 to 25 in the most AI-exposed occupations now sits about 19% below where it would be had it kept pace with their less-exposed peers. Experienced workers show no comparable gap. The researchers point to a distinction every owner should understand. AI is good at reproducing knowledge that’s already written down in manuals, procedures, and textbooks. It’s much weaker with the knowledge people pick up through practice, mentorship, and years of handling whatever the day throws at them. That second kind lives in your people’s heads, and it leaves when they do.
of service firms using AI in the New York Fed’s region laid off workers because of it, while just over a third retrained their people instead. Federal Reserve Bank of New York, Regional Business Surveys (August 2026) →
So what do you do instead? Before any role changes because of AI, I walk owners through four questions, and I want the answers in writing.
First, what does this person know that isn’t written down anywhere? Which customers trust them by name, which workarounds keep a process from breaking, which problems they catch before you ever hear about them. If you can’t answer that, you’re not ready to make a decision about their job.
Second, which part of the job is AI taking, and what do you want this person doing with the hours it frees up? Most owners have a list of things nobody has time for: follow-up calls, training the new hire, fixing the process everyone complains about. Those are the hours AI just handed you.
Third, what training will they get, and who is accountable for it? The firms in the New York Fed survey described everything from basic AI literacy to learning how to check AI output and protect company data, delivered through workshops, outside help, and informal show-and-tell sessions between coworkers. None of that needs a big budget, though it does need someone whose job it is to make it happen.
Fourth, have you rewritten the job description? If the work changed and the written expectations didn’t, you’ve set someone up to fail and then blamed them for it. Expectations and accountability only work when the expectations are current.
And if AI has you thinking about skipping your next entry-level hire, slow down. The Stanford data suggests the adjustment is happening mostly through fewer young hires rather than firings. That feels painless today. But the knowledge your senior people carry has to pass to somebody, and it only passes through practice and mentorship. My advice on hiring hasn’t changed: take your time, be patient, and over-interview. “You need to kiss a lot of frogs in order to find the one prince that you’re looking for.” Stop hiring altogether and one day you’ll have nobody to hand the business to.
I’ve spent 35 years and more than 500 organizations watching owners decide what kind of company they’re building, whether they meant to decide or not. Culture defaults to the personality of whoever’s running the business, and AI is going to make that default louder. If your first move is the cut, your people will learn that fast. If your first move is the training, they’ll learn that too, and they’ll bring you their best work because of it.
If you’re still sorting out where AI fits in your business, start with Most Teams Use AI for One or Two Tasks and Stop There. For the hiring side of this decision, read The Hidden Cost of Always Hiring From Outside. The Inside-Out Method™ begins before anyone buys another tool, with the owner writing down what kind of company they want their people working in.
