No-hire business applications rose 26.8% in AI-exposed industries

Alex Willen runs 12 e-commerce brands and expects them to bring in between $2.5 million and $3 million this year. His headcount is one.

Kristin Ginn runs a similar setup in services. After leaving Microsoft in March, she launched trnsfrmAItn, a consultancy that helps companies adopt AI. Seven months in, she had roughly $250,000 in work either already received or under contract through the end of the year. She has no staff and no regular contractors. CNBC profiled both founders on November 7 alongside new research suggesting they belong to a measurable pattern rather than being two outliers.

That pattern matters most to whoever sells to new small businesses: the bank opening the operating account, the payroll provider, the software vendor counting on a team that grows into more seats. When a company can scale revenue without hiring, the products built around its first employee lose their usual entry point.

Applications that skip the payroll step

The research comes from Liya Palagashvili, a labor economist who directs the Labor Policy Project at George Mason University's Mercatus Center, and it draws on U.S. Census Bureau business formation statistics. She found that in industries with high AI exposure, applications showing no early signs of hiring climbed 26.8%, reaching roughly 281,000 a quarter from about 222,000, when the first quarter of 2026 is compared with the same quarter of 2024. Over that period, applications in those industries with stronger signs of becoming employer businesses dropped 6.4%.

The comparison group is what gives the figure its weight. Palagashvili measured against construction and wholesale trade, where no-hire applications edged down 0.4% during the same period. That gap might reflect something particular to AI-exposed work. Palagashvili warned, however, that her results are preliminary and only descriptive, and that they do not prove AI caused the increase.

Palagashvili describes the findings as preliminary and descriptive, and says they don't establish that AI caused the increase. The Census measure flags applications that lack the usual indicators of near-term payroll or hiring plans. It can't tell whether those businesses stayed one-person firms or hired later.

That gap is a familiar one. A September 2025 Census Bureau Center for Economic Studies working paper by Christopher Goetz, Henry Hyatt, Zachary Kroff, Kristin Sandusky and Martha Stinson noted that the youngest and smallest businesses tied to self-employment are often missing from economic data. The authors built a longitudinal dataset following owners of tens of millions of nonemployer businesses, including their entry, exit and changes in legal form. Applications are the earliest signal available. Tracking what these firms become will take that kind of follow-up data.

Where the payroll money goes instead

Willen's books show what a no-hire firm spends money on instead of wages. Across his brands, he relies on Anthropic's Claude for bookkeeping, customer service, advertising and inventory, and he uses OpenAI image tools for work that photographers and designers used to do. He told CNBC that without AI he would likely need another full-time employee, plus contractors to cover photography, design, bookkeeping and ad management. A designer used to charge $300 to $1,000 for a fresh set of listing images, and a traditional product shoot cost $2,000 to $5,000. Every one of those costs that AI eliminates, he said, adds directly to his profit.

The Lone Trail figure shows AI opening up a sales channel, not just cutting costs. Lone Trail is a hunting-gear brand whose products include binocular harnesses, packs, and other outdoor accessories. Willen said that when he acquired it, the Shopify site brought in roughly 1% of sales. He chose to keep the channel and rebuilt the site using ChatGPT and Claude, with AI also helping to manage Meta advertising. Shopify supplied roughly 30% of the brand's revenue over a recent 30-day period, he said.

For a lender or supplier, having no payroll does not make a firm low risk. For banks underwriting AI-run sellers, channel concentration could be a more useful risk question than headcount.

Solo at launch, maybe not for long

Both can be true. AI may let founders push back the first hire and then help them grow enough to make one. If so, the Mercatus figure describes how businesses start more than how they end up, and products sold to them would need to cover a longer solo phase before a team arrives.

Stanford economist Erik Brynjolfsson told CNBC the common mistake is to treat AI as a personal productivity tool when it's increasingly doing work that used to need a team. He expects a longer tail of strong small producers alongside more power concentrated in the hyperscalers, which could leave medium-sized firms in the weakest position. For vendors whose customer base sits in that middle, his view is worth testing against their own account data.

Banks build for the owner-operator

Some lenders are already designing for this customer. American Banker's John Reosti reported on November 3 that Huntington plans to roll out an AI operating system for time-strapped owner-operators. The initiative, 1866 Studio, launches with eight AI agents, and President Brant Standridge expects more than 60 within a year, with weekly releases of new agents once the program hits its stride. Goldman Sachs is taking a different route, focusing on education and research for small-business clients.

Huntington's plan lines up with Willen's setup, where Claude helps with bookkeeping, customer service, advertising and inventory. CNBC reported that in highly AI-exposed industries, applications for businesses with no early signs of hiring grew 26.8% between the first quarters of 2024 and 2026, while those with stronger indications of eventually becoming employer businesses declined 6.4%. The measure does not show whether those businesses stayed solo, but it suggests that more new small-business customers in those sectors could start out with no payroll. A bank that offers agents to that owner could become part of the toolset that replaces a first hire. That is a different product role from the payroll account it would once have sold.

Banks are also using the same technology on their side of the relationship. At American Banker's Small Business Banking conference on October 27, a session moderated by American Banker AI analyst Larry Cao brought together Teradata's Gary Class, Citadel Federal Credit Union's Michael Desimone and Capitol National Bank's Cristin Reid. They discussed agentic systems that spot cross-sell opportunities, flag at-risk relationships, prepare call briefs and automate follow-up, along with the governance needed to run them responsibly. A separate fireside chat with Thread Bank's Amanda Coleman examined how banks bring accounting, payroll and e-commerce services to small business clients eager to manage cash flow more effectively.

Signals for the next four quarters

The Census series itself will test this reading. Quarters after early 2026 will show whether the gap between no-hire and employer-type applications in AI-exposed industries keeps widening, and whether construction and wholesale stay flat. The measure tracks applications, not what happens after launch, so it will not show on its own whether these applicants stay solo.

The second is Huntington's own target. If 1866 Studio grows from eight agents to more than 60 within a year, as Standridge expects, owner-operators will be able to choose between a bank's agents and the Claude or ChatGPT subscriptions they already use. Whichever they pick will show small-business banking teams which product actually replaced that first hire.