Skip to content
All posts

How to check what AI apps are telling people about you

A repeatable twenty-minute audit across four apps, and a way of recording it so that doing it again in three months actually tells you something.

Most owners have asked an AI app about their business once, got an unflattering answer, felt slightly worse, and learned nothing they could act on. The difference between that and a useful audit is about twenty minutes and a sheet of paper, and it is worth doing properly because the result determines what you fix.

Start in a private window

If you have chatted about your own business before, the app may remember, and it will tell you a version of things shaped by everything you have already said to it. Open a private window, or start a fresh chat with memory turned off, so that you see roughly what a stranger sees. This single step changes the result more than anything else here, and skipping it makes the whole exercise worthless.

Step one: ask the way a customer would

Ask ChatGPT, Gemini, Claude and Perplexity the same thing, phrased the way somebody who does not know you would phrase it: who is a good [your trade] in [your town]. Write down every business each of them names, in the order they come.

Then run it a second time in each app. You will often get a different list, and that is normal rather than a fault. Knowing how much the answer moves around on its own is what stops you over-reacting to a single bad result later.

Step two: ask about yourself by name

Now ask each app what it knows about your business, and then push on the specifics that decide whether somebody books: your opening hours, the areas you cover, roughly what you charge, and whether you take on the particular job you get asked about most.

Record three columns as you go. What it got right, what it got wrong, and what it would not commit to at all. That third column is usually the longest and it is almost always the most useful, because the things an app hedges on are the things your website never actually said.

Step three: ask the question that decides the job

Every trade has one question that separates a real enquiry from a wasted morning. Do you do emergency call-outs. Do you work on commercial units as well as houses. Do you come out to that village fifteen miles away. Do you take that insurer.

Ask the apps your version of it. If they cannot answer, then a customer reading your website cannot answer it either, and you have found something worth fixing regardless of what you think about any of this.

Reading what you got back

There are three outcomes and they need three different responses, which is why people who skip the recording step often fix the wrong thing.

If you are missing entirely, the job is to publish basic facts in plain text where they can be found: hours, area, and what you actually do. If the description is wrong, the job is detective work, because somewhere there is an old listing or a forgotten page that still says so, and it needs correcting or deleting. If the answer is vague, the job is turning soft language into statements, because "we pride ourselves on quality" is not a fact anybody can repeat.

Do it again in three months

A single reading is worth very little on its own, because the comparison is where the value sits. Keep the sheet, put a reminder in the calendar, and repeat it quarterly. Ten minutes, four times a year, and you will know whether what you changed made any difference, which is more than most people selling services in this area can tell you. If the first run comes back badly and you want the reasons before you change anything, start with how AI apps decide who to recommend.

Written by the team at hailera — we help small businesses answer customer questions without being there.