Stax Says 80 Percent Of Small Firms Replaced Experts With AI. Self-Reported, Naturally. The Distinction Matters.
A Stax survey reports that 80 percent of small businesses use AI to replace expert work, with 80 percent expecting increased AI use within six months. The survey also claims 96 percent say AI speeds decisions, 95 percent say it increases confidence, and 94 percent say it improves customer service. The source itself notes these are self-reported perceptions, not measured outcomes like revenue growth or error rates.
The concept here is the difference between perceived and measured utility. Self-reported confidence tells you about feelings. It tells you nothing about whether decisions actually improved. The mechanism is the Dunning-Kruger amplifier. AI makes people feel more competent. Feeling competent and being competent are not the same thing. Small businesses replacing accountants and HR consultants with language models should track actual error rates, not just satisfaction.
Stax, a payment processing company, surveyed small businesses and found 80 percent substituting AI for expert advisory work. The article notes that local accountants, tax preparers, designers, and HR consultants currently serve the same ecosystem that banks and payment firms want to automate.
- Pick one business decision you recently made, such as a pricing change or a vendor choice. Write down what you decided and why.
- Ask ChatGPT or Claude to analyze that same decision. Give it the same context you had at the time. Compare its reasoning to your actual reasoning.
- Document where the AI agreed with you, where it disagreed, and where it introduced information you lacked. This is your first data point on whether AI actually improves your decisions or merely makes you feel better about them.