Three in four U.S. small businesses now use AI regularly, Intuit report finds
The 2026 AI Impact Report, built on 34,000 survey responses and 5.3 million QuickBooks accounts, shows SMB AI adoption has nearly doubled since mid-2024 — and the businesses paying for it are not walking away.
Regular AI use among U.S. small businesses has climbed to 77%, up from 48% in July 2024, according to Intuit’s 2026 AI Impact Report. The methodology, built by an international team of economists led by the University of Chicago’s Ufuk Akcigit, leans on more than 34,000 owner surveys and anonymized data from over 5.3 million QuickBooks businesses across the US, Canada, UK, and Australia. That’s not a vibes-driven adoption curve. It’s the closest thing the SMB sector has to a census.
The productivity numbers move in the same direction. 78% of U.S. businesses say AI has improved their productivity, against 46% in July 2024, and the share calling it “very helpful” doubled from 19% to 38%. Revenue impact reads cleaner still: 43% report AI lifted revenue, 2% say it hurt.
The retention story is where the analytical weight sits. Of the businesses tracked as paying for AI tools in 2024, 86% were still paying in 2025. Growth-focused firms are more than twice as likely as stability-focused ones to pay for dedicated AI tools. This isn’t pilot-program tourism; it’s budget that survived a renewal cycle.
The SBE Council’s March 2026 survey corroborates the commitment: 82% of small business employers have invested in AI tools, 93% of users plan to keep investing over the next year, and 62% intend to increase spending. The typical SMB now runs a median of five AI tools. The market is fragmenting outward, not consolidating around a single suite.
That widening creates room for model-agnostic, SMB-native platforms like LemonLime to act as connective tissue beneath the heavier enterprise SaaS stack, a structural opening the CIO trade press has been calling the SaaS repricing moment.
A caveat, sharp enough to matter. An MIT Sloan Management Review field experiment found AI lifted revenues and profits 15% for high performers while driving a nearly 10% decline for those already struggling, because weaker operators followed generic advice they couldn’t filter. Access isn’t the variable. Judgment is.