Sixty-seven percent of small and mid-sized businesses that use AI have built at least one AI tool now in regular use, according to Pax8’s Q3 2026 SMB AI Pulse Report, released October 12. The survey of 1,150 business leaders also found that 35% discovered AI running without approval and 49% received AI bills higher than expected.
Key Takeaways
- In the United States, 81% of AI-engaged small and mid-sized businesses have built AI solutions now in regular use, compared with 67% across all four markets surveyed.
- Only 36% of U.S. businesses building AI have a formal, documented review process for the tools they create.
- 57% of U.S. businesses that paid for AI in the past year received at least one bill or usage charge higher than expected.
- 55% of decision-makers personally spend three or more hours a week reviewing or fact-checking AI-generated work.
- A separate New York Fed analysis found AI-using small firms had 14 percentage points higher net expected revenue growth than non-users, after controlling for firm characteristics.
- 70% of surveyed businesses said they would pay at least 6% more for greater control over how their data is stored and used.
Small Businesses Have Moved From Using AI to Building It
The survey’s main finding is a change in how small businesses use AI. Across the United States, United Kingdom, Germany, and Sweden, 67% of AI-engaged businesses have built one or more AI solutions now in regular use, and another 19% are prototyping. Only 13% rely solely on off-the-shelf tools.
The building takes several forms. Among businesses building AI, 52% are creating automated workflows that connect AI to other systems such as finance and operations software, 46% are building or testing AI agents that complete multi-step tasks on their own, and 39% have built custom integrations using an AI provider’s application programming interface (API).
That shift brings new responsibilities. A business that connects AI to its accounting system or deploys an agent to handle customer requests is now responsible for maintaining, securing, and paying for that system, work that software vendors used to handle. For a company with 20 employees and no IT department, that is a real operational commitment.
The pressure to move quickly is strong. 66% of respondents said their business will not stay competitive without AI in the next three years, and 65% said they need to act immediately or within six months.
Shadow AI Shows a Governance Gap
The survey found oversight has not kept up with adoption. 35% of AI-engaged businesses found “shadow AI” in the past six months, meaning tools, workflows, or agents running without official approval.
Formal review processes vary widely. The share of businesses with documented review of internally built AI ranges from 23% in the United Kingdom to 45% in Sweden. The U.S. gap is the most notable: 81% of U.S. AI-engaged businesses have built tools now in use, but only 36% of builders have formal, documented review, and another 41% rely on consistent but undocumented review.
That matters for small firms because data security or privacy exposure was the top-cited risk of employee-built AI in every market, at 30% to 35% of respondents. An unapproved tool connected to customer data or payment systems is a liability most small businesses can’t easily absorb.
Usage-Based Pricing Is Producing Unexpected Bills
Cost control is the second weak spot. Among businesses that paid for AI in the past year, 49% received at least one bill higher than expected. The figure was 57% in the United States and 65% in Sweden.
Usage-based charges were the hardest to forecast, and they grow as companies add more agents and automated workflows. For businesses that budget monthly, that unpredictability matters. A practical step is to set spending limits and usage alerts on every AI account and assign one person to review AI charges each month, the same way many firms already manage cloud or advertising spending.
Time Savings Come With Review Time
The productivity results are mixed. 69% of AI-engaged businesses said AI freed up at least some capacity. But 55% of decision-makers spend three or more hours a week reviewing, editing, or fact-checking AI-generated work, and 7% spend ten hours or more.
For an owner billing their time or managing a small team, three hours a week is close to a full workday each month. The net benefit depends on how much time AI saves compared with how much time goes into checking its output, and the survey suggests many businesses are not yet measuring both.
Federal Reserve Data Shows AI Users Expect Stronger Growth
Independent data from the Federal Reserve Bank of New York supports the growth side of the picture. Economists Will Aarons and Asani Sarkar analyzed the 2025 Small Business Credit Survey, fielded September through November 2025 among employer firms. Forty-six percent of those firms reported using AI, and another 15% planned to adopt it within a year.
AI users were more optimistic than non-users. Their net expectation for higher employment over the next 12 months was 33 percentage points, compared with 15 for non-users. For revenue, the gap was 48 points versus 21. After controlling for firm and owner characteristics and past performance, AI users still showed a 14-point advantage in expected revenue growth and a 10-point advantage in expected employment growth.
The New York Fed also found that the relationship was strongest among firms in weaker financial condition. Among owners reporting fair or poor finances, AI users had 24 points higher net revenue expectations than non-users. The researchers suggest AI may be helping some small businesses overcome specific technical limits rather than simply reflecting general optimism. They also note that whether these firms actually achieve the growth they expect has not yet been measured.
Businesses That Use AI Extensively Feel More Competitive
In all four markets in the Pax8 survey, businesses that use AI extensively were more likely to say they feel ahead of competitors than businesses still experimenting, at 63% versus 40% in the United States. The U.S. was also the only market where fear of falling behind outweighed fear of making costly AI mistakes.
That combination, a strong push to adopt along with weak oversight and surprise costs, sums up where many small businesses are now. The practical steps are straightforward: keep a list of every AI tool in use, require approval before tools are connected to company data, set spending limits, and track the time spent checking AI’s work alongside the time it saves.
FAQs
What percentage of small businesses are building their own AI tools?
Pax8’s Q3 2026 SMB AI Pulse Report found that 67% of AI-engaged small and mid-sized businesses have built at least one AI solution now in regular use, rising to 81% in the United States.
What is shadow AI in small businesses?
Shadow AI refers to AI tools, workflows, or agents running inside a business without official approval. The Pax8 survey found 35% of AI-engaged businesses discovered shadow AI in the past six months.
Why are small businesses getting unexpected AI bills?
Usage-based pricing is the main cause. Charges rise as businesses add AI agents and automated workflows, and 57% of U.S. businesses that paid for AI received a bill higher than expected.
Does AI adoption help small business revenue?
New York Fed research found AI-using small firms expect stronger revenue growth than non-users, by 14 percentage points after controls. Actual revenue outcomes have not yet been measured.
How was the Pax8 SMB AI Pulse survey conducted?
The online survey covered 1,150 decision-makers at companies with 5 to 499 employees in the U.S., U.K., Germany, and Sweden. The margin of error is about ±2.9 points overall and ±4.9 points for the U.S.
