Small businesses that use artificial intelligence (AI) are more optimistic about hiring and revenue over the next year than those that do not, according to an analysis published October 8 on the Federal Reserve Bank of New York’s Liberty Street Economics blog. The authors stress that the findings describe owners’ expectations, not results.

The post, by Will Aarons, a research analyst, and Asani Sarkar, a financial research advisor, both in the New York Fed’s Research and Statistics Group, draws on the 2025 Small Business Credit Survey (SBCS), which the Federal Reserve Banks fielded in the fall of 2025. The AI questions covered 5,248 employer firms, with responses weighted to match the national population of employer firms.

Who is using AI, and for what

In the survey, 46% of employer firms said their business or employees were using AI tools, and another 15% planned to start within 12 months. The Fed’s 2026 Report on Employer Firms, published in March from 6,525 survey responses, found that a third of firms had no plans to use AI. Among those, more than half said AI was not applicable to their business, and 30% said they preferred not to use it.

Among users, the report found about half were still experimenting, 44% had partially integrated AI into their business processes and 7% had fully integrated it. The most common uses were writing or marketing (83%), individual productivity (61%) and planning or analysis (51%). Users most often cited accuracy (46%) and adapting tools to business needs (43%) as challenges.

Reported benefits were uneven. According to the report, 71% of users saw higher productivity and 39% saw better quality of goods and services, but only 31% said AI had increased their sales. The New York Fed post notes that 77% of firms reported no change in labor costs from AI use.

The expectations gap

The New York Fed analysis measured “net” expectations: the share of firms expecting an increase minus the share expecting a decrease. Among firms with at least one employee, the net share expecting higher employment was 33 percentage points for AI users and 15 points for non-users. For revenue, the figures were 48 points for users and 21 for non-users.

Part of that gap reflects other differences between the firms. Controlling for firm and owner characteristics and location cut the employment advantage for AI users to 13 points, and adding past employment changes cut it to 10 points. On revenue, AI users kept a 14-point advantage after controls that included past revenue changes. The authors also found that younger firms, more profitable firms and firms with more than 10 employees were more optimistic about hiring.

How far a firm had gone with AI made little difference. After controls, users that had fully or partially integrated AI, or that called it important to production, were as likely to expect higher employment as firms still experimenting with it.

Why it matters

For software vendors selling AI features to small merchants, the survey describes a market in which 46% of employer firms already use AI and another 15% intend to, with marketing and writing the leading use and accuracy the leading complaint. It also shows a gap between perceived productivity gains, reported by 71% of users, and sales gains, reported by 31%.

The data do not show small firms using AI to cut payroll so far. Most users reported no change in labor costs, and AI users expected to add workers at a higher net rate than non-users. The link between AI use and optimism was strongest among owners under financial strain: for those rating their condition fair or poor, the employment gap between users and non-users was 21 points, against 10 points for owners in good or better condition. The revenue gaps were 24 and 15 points. Among AI users, firms reporting technology-related operational challenges had net revenue expectations 16 points higher than firms without them.

The caveats

The study measures expectations, not outcomes, and shows an association rather than cause and effect. The authors used owners’ self-rated financial condition as a proxy for general optimism, to test whether optimistic owners are simply more likely both to adopt AI and to expect growth. Because the gap was larger among owners in weaker financial condition, they argue the link is unlikely to be spurious. They also wrote that “Whether AI-adopting firms actually experience the stronger revenue and employment outcomes they anticipate is a subject for future research.”

The employment analysis covers only firms with at least one employee, and the post does not break results down by industry, so it says nothing specific about retailers or online sellers. The post states that the views are the authors’ own and do not necessarily reflect the position of the New York Fed or the Federal Reserve System.