The short version: small businesses using AI are hiring more people than comparable businesses that are not, according to two independent datasets published in the past seven weeks. Gusto’s payroll records show AI adopters ended their first year with about 7 percent more staff than similar non-adopters, and the Federal Reserve Bank of New York found AI users far more likely to expect to add employees. The part nobody put side by side: both datasets find the effect concentrated where an owner is stretched thinnest, in shops under ten people and owners in fair or poor financial shape, and both find that everyday use counts as much as deep integration.
That second finding matters more than the headline number. In Gusto’s data, a business with fewer than ten employees grew its team about 10 percent more than a comparable one that knew about AI and chose not to use it. Businesses with 10 to 24 employees showed no measurable headcount change at all. If you run a four-van shop and have been told AI is coming for your crew’s jobs, the best evidence available this fall points the other way, and points hardest at you.
What the small business AI hiring data actually shows
Three sources landed between September 10 and October 8, 2026. They measure different things, which is why they are worth reading together rather than one at a time.
Gusto’s report, by economists Nich Tremper and Tom Bowen, is the only one built on what businesses actually did. It follows 2,262 Gusto payroll customers who answered the company’s April 2025 survey: 1,593 that said they used generative AI and 669 that knew about AI but had not adopted it. Then it counts heads on payroll, quarter by quarter, for a year after each business started using AI.
The New York Fed analysis, by Will Aarons and Asani Sarkar, uses the 2025 Small Business Credit Survey, fielded September to November 2025, covering 5,248 employer firms weighted to match the national population. It measures what owners expect over the next year, not what happened.
The Small Business and Entrepreneurship Council’s October check-up polled 516 owners with 2 to 99 employees between September 11 and 19, 2026. It is the freshest read on sentiment, and the smallest sample.
| Source | What it measures | Businesses | Headline finding |
|---|---|---|---|
| Gusto, Sept. 10, 2026 | Actual payroll headcount, four quarters after adopting AI | 2,262 | About 7% more staff than comparable non-adopters; about 10% under ten employees |
| New York Fed, Oct. 8, 2026 | Owners’ hiring and revenue expectations for the next year | 5,248 | Net hiring expectation +33 points for AI users, +15 for non-users |
| SBE Council, Oct. 8, 2026 | Owner sentiment, AI agent use | 516 | 58% use AI agents; 16% say AI eliminates more jobs than it creates |
Read across the rows and the pattern holds from three directions. The owners who use AI expect to hire, the payroll records show they did hire, and only one owner in six believes AI is a net job destroyer.
Are small businesses using AI really adding staff, or just optimists?
This is the obvious objection, and both research teams went after it directly.
The New York Fed’s raw gap is 18 points: a net 33 percent of AI users expect to add employees, against a net 15 percent of non-users. Once the authors control for firm size, owner age, profitability and location, the gap shrinks to 13 points. Add each firm’s employment change over the previous year, so a business that was already growing does not get credit for it, and 10 points remain. Then they used the owner’s self-reported financial condition as a stand-in for general optimism. The gap held. Their conclusion, in the authors’ words: “This result is not due to optimistic owners being more likely to use AI.”
Gusto’s method is stricter, because it is not asking anyone what they think. Its economists used an event-study design built for businesses that adopt at different times, comparing each adopter against firms that had not yet adopted or never did, starting from the quarter before adoption. Before adoption, the two groups’ employment lines were flat and parallel, which is the check that they were comparable to begin with. After adoption, the AI users pulled away gradually: about 4.5 percent ahead early on, 6.8 percent by the end of the fourth quarter. A one-time hire would show up as a single jump. A steady climb looks like a business finding more work to staff.
Gusto also checked wages and found no matching change, which rules out the gap being a payroll quirk from giving everyone a raise. The extra payroll is extra people.
Why the smallest shops gain the most
Here is where the two datasets agree on something neither one said about the other.
In Gusto’s numbers, the gain sits almost entirely in businesses with fewer than ten employees: about 10 percent more headcount than comparable non-adopters. Gusto’s own illustration is the average nine-person business, which within a year of adopting AI has, in effect, onboarded one more employee. At 10 to 24 employees, Gusto measured no headcount change at all.
In the New York Fed’s numbers, the gap is widest for owners who rated their finances fair or poor: 21 points on hiring expectations, against 10 points for owners in good or better shape. The same split shows up on revenue, 24 points against 15.
Different measures, same shape. AI is associated with the biggest difference where the owner is the bottleneck. In a six-person shop the owner is also the scheduler, the estimator, the bookkeeper and the person who answers the phone at 9 p.m. Gusto’s authors make this argument themselves: in the smallest businesses the owner’s hours are the scarcest resource, and they go to overhead. A 20-person business usually already has an office manager absorbing that work, so the time AI frees up shows up somewhere other than a new hire.
The arithmetic for your own shop is worth doing honestly, because 10 percent is an average across hundreds of businesses, not a promise to any one of them. For a nine-person business, 10 percent is roughly one person. For a four-person business it is 0.4 of a person, which in practice means that across ten four-person shops using AI, about four more people were hired than across ten that were not. Nobody hires 0.4 of a technician. What the number describes is how much more often a small shop using AI reaches the point where adding someone makes sense.
Who they hired: do-ers, not AI engineers
If AI hiring meant small firms bringing on a data scientist, it would not mean much for a plumbing contractor. Gusto looked at the job titles, and the new hires were the people who do the work customers pay for: therapists and care staff at health and social assistance businesses, teachers, technicians and cooks elsewhere, plus some front-desk and administrative support.
The gains were not confined to tech or professional services either. Goods-producing businesses such as construction and manufacturing, and personal services such as restaurants and salons, saw growth similar to accountants and tech firms. The standout was small community-services businesses, mostly health care and social assistance under ten employees, at 19 percent more headcount. Gusto points to examples such as more dental hygienists at dental practices and more therapists at mental health clinics, and suggests the mechanism is documentation: when paperwork takes less of a clinician’s day, there is room for more patient appointments, and eventually for another clinician.
This matches what an earlier study found at a much larger scale. A 41-country analysis we covered in September found that companies adopting AI grew their headcount, with most of the new seats going to experienced workers. That paper was about multinational employers. The new data is about American Main Street, and it says the same thing more loudly at the small end.
You do not need to go all-in to see it
The finding that should lower the temperature for most owners is about depth of use, and again both datasets agree.
Gusto reports that businesses using AI for everyday tasks saw the same hiring gains as those that had built it deep into their operations. Its own examples are modest: a solo operator using AI to draft proposals, a dental office using it to speed up paperwork. The New York Fed found the same thing in its own terms: among AI users, owners who said they were only experimenting were as likely to show the higher hiring expectations as owners who had integrated AI partially or fully, after controls.
That is a useful corrective to the way AI gets sold to small business, which often implies the payoff waits on the far side of a large project. The research suggests the payoff begins with ordinary tasks that eat an owner’s evening: the estimate write-up, the follow-up email, the job notes, the intake paperwork. A tool like Claude Duet, which drafts invoice follow-ups and preps payroll for review, sits squarely in that category, and so does the chat assistant most owners already pay for.
What the research does not prove
Good news deserves the same scrutiny as bad, so here are the limits, stated by the researchers themselves.
Both studies describe associations, not proof that AI caused the hiring. Businesses choose whether and when to adopt, and early adopters may differ in ways nobody measured. Gusto says so plainly in its methodology, and the New York Fed authors write that whether AI-using firms actually achieve the growth they expect is a question for future research.
Gusto’s sample is its own customers who answered a survey, not a random draw of American businesses, and Gusto sells payroll software to small businesses, so it has an interest in an encouraging story. That is a reason to read its methods closely, and they hold up: a named comparison group, a pre-adoption trend check, and robustness tests that change the comparison group and restrict the sample to established firms. The New York Fed survey, which has no product to sell, independently points the same direction.
And AI is not a growth machine on its own. In the New York Fed’s survey, 77 percent of AI users reported no change in labor costs from AI, and only 31 percent said it had increased their sales. Most owners are getting time back, not a windfall. The hiring shows up when that time goes into more customer work.
What to do with this as an owner
Three practical takeaways, none of which requires buying anything new.
Start with the work that keeps you at the desk. The research points at overhead in the smallest shops: scheduling, estimates, follow-ups, documentation. Pick the one task that most often keeps you off the job or up at night, and try an AI assistant on that alone for two weeks.
Watch where the hours go. The businesses in Gusto’s data did not bank the freed time; they turned it into more customer work and, eventually, another hire. If AI saves you five hours a week and they disappear into more email, nothing changes. Decide in advance what those hours are for. If you want a number to start from, NucoAI, our sister company, built a free AI efficiency calculator that estimates how many usable hours your current AI spend buys each month and what each of those hours really costs once staff time is counted. No signup is needed to see the result.
Plan the hire before you need it. If the pattern holds for your shop, the moment arrives sooner than you expect. Our playbook on building a hiring process backward from the start date covers the clocks that start running once you decide, including the point where letting AI screen applicants becomes a regulated act. For the broader worries owners raise about AI, from cost to liability, see the concerns worth having.
Frequently Asked Questions
Are small businesses that use AI hiring more people?
Yes, on the best evidence available in 2026. Gusto’s payroll data on 2,262 businesses found AI adopters ended their first year with about 7 percent more staff than comparable businesses that knew about AI but did not use it, and a New York Fed survey of 5,248 employer firms found AI users much more likely to expect to add employees. Both studies show association rather than proof of cause.
Which small businesses gain the most from AI in hiring?
The smallest and most stretched. Gusto found businesses under ten employees grew their teams about 10 percent more than comparable non-adopters, while businesses with 10 to 24 employees showed no measurable change. The New York Fed found the gap widest among owners who rated their finances fair or poor.
Do you need a big AI project to see these gains?
No. Gusto reported that businesses using AI for everyday tasks saw the same hiring gains as those that had built it deep into their operations, and the New York Fed found that owners who were only experimenting showed the same higher hiring expectations as those who had integrated AI.
What kinds of jobs did AI-using small businesses add?
Mostly hands-on, customer-facing roles, according to Gusto’s review of job titles: therapists and care staff, dental hygienists, teachers, technicians and cooks, plus some front-desk and administrative support. The gains appeared across construction, manufacturing, restaurants, salons and professional services, not just in tech.
If you have started using AI in your business, where did the time it saved actually go? Tell us in the comments.
