99% of Small Businesses Use AI. 40% Still Feel Unprepared

Your business already runs on AI. So why does nothing feel easier?

That is the question raised by a new TD Bank survey of 200 U.S. small business owners, conducted by Wakefield Research in late August and released this week. Nearly every respondent, 99%, says they use AI in some capacity. More than three quarters say their use has increased in the last six months. And yet 40% say AI adoption is one of the areas where they feel unprepared, right behind technology and operations generally at 43%. (Source: TD Small Business Year-End Outlook & Preparedness Survey, via Business Wire, Oct 5, 2026.)

Read that again: almost universal adoption, and almost half feel unprepared. The tools are in the building. The confidence is not.

The problem was never the tools

When owners say what they use AI for, the answers are refreshingly practical: saving time (57%), better cash-flow decisions (55%), and increasing sales (52%). Nobody is asking for revolutionary AI employees. They are asking for Tuesday afternoons back.

The reason Tuesday still hurts is that AI landed in pieces. A chatbot here, a scheduler there, a content tool somewhere else, none of them talking to each other. The inquiry gets answered by one tool but never logged in the system. The quote gets built in one place and invoiced in another. Every handoff between tools is still a human with copy and paste. Businesses did not fail to adopt AI. They adopted five of them, and the work between them is still manual.

That is exactly the gap the survey is measuring: 99% adoption of AI, 40% unpreparedness about AI. Both numbers are true at the same time, because the missing piece was never another tool. It is the workflow between the tools you already have.

What "prepared" actually looks like

Imagine an inquiry coming in after hours. The request is understood, the right follow-up questions are asked, your business rules are applied, your existing software is updated, scheduling is checked, the customer gets an answer, a follow-up goes out if they go quiet, and anything unusual lands on a staff member's desk in the morning.

Nothing in that picture requires new software. It requires the software you already use to finally work together, with the repetitive steps handled automatically and the exceptions handled by people. That is what prepared looks like. Not more AI, but the AI you have doing the whole job.

Honesty is becoming the moat

There is a second development worth noting. Last week the FTC confirmed it has opened an inquiry into OpenAI, Anthropic, and other AI companies over consumer-facing product risks (reported by CNBC on Sept 30). No wrongdoing has been alleged, and no documents have been published. But the direction of travel is clear: if the AI labs themselves may have to show that their marketing matches reality, vendors selling AI magic will be judged by the same standard.

That is why we keep our positioning plain. We do not promise specific integrations we have not built, we do not call anything fully autonomous that is not, and we do not publish pricing for work we have not scoped. Honest automation is not just a nicer pitch. It is the one that survives scrutiny.

The survey tells us businesses are confident and spending again, 85% expect stronger revenue through year-end. The ones who close the gap between their AI tools will feel it first. If you want to know where your biggest gaps are, that is exactly what a workflow review maps out.