How to Check Whether ChatGPT Recommends Your Business

Why you get a different answer every time

The first thing that unsettles people doing this check is inconsistency. You ask the same question twice and get two different lists. Nothing is broken; this is how these systems work, and understanding why saves you from drawing the wrong conclusion from one test.

Three things are moving at once. The models are non-deterministic by design, so the same prompt can take a different path on each run. Some tools search the live web for each answer while others lean on what they absorbed in training, which means one has today's information and the other has last year's. And personalisation matters: memory, custom instructions, your location and your chat history all shape what you see. Logged into your own account, you are a uniquely bad test subject for your own business.

So test properly. Use a logged-out or private window, turn memory off where you can, and run each question three or four times rather than once. Vary the phrasing the way real people do, and check from the location your customers are actually in. What you are looking for is a pattern across repetitions, not a verdict from a single answer.

What to fix first when you are invisible

If those checks turn up nothing, the instinct is usually to write more blog posts. That is rarely the bottleneck. Work through these in order, because the early items gate the later ones.

Make your basic facts consistent everywhere. Name, address, phone number, what you actually do, the areas you serve. If your website says one thing, your Google Business Profile another and an old directory listing a third, you have given every engine a reason to trust none of them. This is dull, fast work and it is the most common single cause of invisibility.

Say plainly what you sell and to whom. A surprising number of business websites never state it in a sentence a machine could lift. "We deliver in-person AI training for small businesses across Oxfordshire" is extractable. "Transforming tomorrow, together" is not.

Get mentioned where the engines already look. This is the step people skip and it matters more than your own site. Industry directories, professional bodies, local business organisations, review platforms, supplier and partner listings, genuine press coverage. Models weight corroboration from sources they did not have to take your word for.

Then build the content. Pages that answer the actual questions a buyer asks, with a direct answer near the top and specifics rather than adjectives.

The surfaces that matter more than your own website

Here is the uncomfortable part. When an AI recommends businesses, it is frequently synthesising from third-party sources rather than reading company websites, because a company website is a self-interested source and the model knows it. Which means a chunk of your AI visibility sits on property you do not own.

In practice the recurring sources are review platforms, established directories in your sector, membership and accreditation registers, local and trade press, and community discussion where people ask for recommendations by name. None of that is something you can optimise in an afternoon, and none of it responds to keyword tricks. What it responds to is being a real, verifiable, well-regarded business with a consistent footprint — which is a slower game, but also one that does not evaporate the next time a model is retrained.

A word of caution, since it needs saying: do not attempt to manufacture this with fake reviews or paid-for mentions dressed as editorial. Review platforms are aggressive about detection, the reputational downside is severe, and models increasingly discount sources that look astroturfed. The shortcut is worse than the long way round.

Does ChatGPT Recommend You? – frequently asked questions