The Stealth BriefHiring Economics

Two-Thirds of Small Businesses Use AI. Hiring Is Still Their Hardest Problem.

6 min readStealth Agents Editorial2 sources cited

Small business owners have adopted AI faster than almost anyone predicted. They still cannot hire. Those two facts sit next to each other in Gusto's 2026 State of Small Business report, and the gap between them is the useful part.

The news

Gusto surveyed 1,731 owners and operators of U.S. businesses with fewer than 25 employees between 17 August and 7 September 2026. The margin of error is +/- 3.4 points.

The headline numbers:

  • 67% use AI tools. That breaks into 19% experimenting, 28% using AI for at least some tasks, 15% using it regularly across multiple functions, and 5% saying it is built into every part of the business where possible.
  • Adoption is highest in professional services, where 27% use AI across multiple functions or the whole business, against 16% in personal services and 13% in community services.
  • 25% of businesses with employees now have an AI policy, up from 15% last year.
  • 61% are optimistic about their own business, flat year over year. Only 18% are optimistic about the U.S. economy, down from 29%, and 59% describe themselves as pessimistic.
  • 86% say their business is meeting or exceeding the expectations they had when they started it.

On hiring: 26% of owners tried to hire in the past year and struggled to find qualified people they could afford. Among those, 64% said candidates lacked the skills the job required. The smallest businesses were more likely to say they could not afford what qualified candidates wanted; larger small businesses more often ran into scheduling conflicts.

Costs are the other pressure. Fifty-nine percent say changing input prices hurt them, 40% say rising interest rates hurt, and 52% raised prices in the past year, most by up to 10%. Half plan to raise prices again next year.

"Small business owners are telling us two things at once," said Nich Tremper, Senior Economist and Head of Gusto Insights. "Eighty-six percent say their businesses are meeting or exceeding expectations, yet more owners than a year ago say the economy is uncertain. Higher costs and expensive borrowing make it hard to plan ahead."

Why it matters

The AI story has been framed as a substitution story: tools replace people. This survey describes a different pattern. Owners are adopting AI and still failing to fill roles, which suggests AI is absorbing tasks, not whole jobs.

Look at what owners are comfortable automating. About 60% of AI users are comfortable letting AI draft marketing content with limited oversight, against 27% of non-users. They are far less comfortable handing it data analysis or direct customer questions. That is a task-level division of labor, not a headcount decision.

Meanwhile the hiring failure is specific: skills and affordability. Not "no applicants." Sixty-four percent of struggling employers said candidates lacked the required skills.

That is the gap a scoped offshore role fills. The work that needs judgment, tool fluency and follow-through does not disappear because a model can draft a paragraph.

What to do about it

1. Separate the task list before you buy another tool. Mark each task automate, delegate or keep. If most of a role is "keep," an AI subscription will not solve the hiring problem.

2. Stop writing job posts for unicorns. The skills gap is often a scoping gap. One owner hiring for "marketing generalist who also does ops and reporting" is competing for a candidate who barely exists at an SMB budget.

3. If you cannot afford the local market, price the alternative honestly. A domestic part-time hire at $30–$45 an hour for 20 hours is roughly $2,400–$3,600 a month before overhead. That is the number to compare against a dedicated offshore seat, not an hourly rate in isolation.

4. Write the AI policy before you need it. A quarter of employers now have one. The minimum viable version: which tools are approved, what data cannot go into them, and who reviews output before it reaches a customer.

5. Keep a human owner for exceptions. The survey shows owners trust AI with drafting and distrust it with analysis and customer questions. That instinct is correct. Assign the exceptions to a person and document the escalation path.

The Stealth take

The clients who get the most from an assistant are the ones who treat AI as a first draft layer and a person as the final mile. They use tools to compress the writing and research, then have someone accountable for the judgment, the client relationship and the exception.

That combination is also the cheapest way to close a skills gap. You are not hiring a unicorn who does everything. You are hiring a person who owns a defined set of tasks and uses the tools you already pay for.

Sources

Related: what a VA costs versus an in-house hire · how to write a role brief

Sources

← All briefs