AI for small business: what actually works
Fewer than one in five small businesses use AI, and the ones that do often bought the wrong thing. From systems running in real clinics and shops, here is where AI earns its place and how to tell a useful project from an expensive one before you pay.
Most of what you read about AI for small business is written by people selling it. I build these systems too, so treat this with the same caution. But I can tell you what has worked in the businesses I run systems for.
Where most small businesses actually are
The US Census Bureau tracks this every two weeks. From December 2025 to May 2026, 17 to 20 percent of businesses said they used AI. Among firms with fewer than 20 employees it stayed under 20 percent, with no real change in six months. Firms with 250 or more employees were at 37 percent.
So if you have not started, you are not late. And the gap is less about technology than about having a clear job for AI to do.
Where AI genuinely helps
Reading documents and filing them. This is the most valuable thing I have built with AI. A clinic receives lab results, imaging reports and letters by email every day. AI reads each one, finds the patient's name and date of birth, and suggests where it belongs. In a busy practice that was hours of a person's week, and the AI does the first pass in seconds.
Turning photos of paper into usable records. Photograph a handwritten form and AI can read it, including a lot of messy handwriting. One practice digitised nearly a hundred patient folders this way from phone photos in a single run.
Answering common questions from your own material. A website assistant that answers from your price list, your opening hours and your preparation instructions, takes the person's details and hands over to a human when it is unsure. It has to be limited to your own material, not the whole internet.
Drafting, not sending. Replies to reviews, follow-up messages, a first draft of a policy. A person reads it before it goes anywhere.
Summarising. Turning a long email chain or a call recording into the five things that matter. Useful, low risk, easy to check.
How to tell a useful AI project from an expensive one
Ask these before you buy anything:
- What exact job does it do? "Saves time" is not a job. "Files lab results onto the right patient file" is.
- What happens when it is wrong? If the answer is "it won't be", walk away. Every AI makes mistakes. A good system makes them visible and sends them to a person.
- Who checks it, and how often? Somebody on your team needs to look at what it did, especially in the first month.
- Where does your data go? Customer and patient information is sensitive. Know which company processes it and whether it is used to train their models. I covered this in client data and AI.
- What does it cost every month, not just to set up? AI tools charge by use. Ask for a realistic monthly figure based on your volume.
A good first AI project
If you want to try AI without much risk, choose something where:
- the job happens many times a week,
- a mistake is easy to spot and easy to fix,
- a person still makes the final decision.
Drafting replies to Google reviews fits. Summarising your day's enquiries fits. Reading incoming documents and suggesting where they go, with a person confirming, fits.
Letting AI book appointments directly into a diary, answer clinical questions, or quote prices on its own does not fit as a first project.
The order that works
AI is the last layer, not the first. Before it can help, your business needs one place where customer records live, your prices written down, and your common answers agreed. I go through that order in how to automate your small business. AI added on top of a mess produces a faster mess.
If you want help working out where AI would actually save you time, I do a free audit.
Want more patients without guessing where they leak? Get your free clinic growth audit