Why ChatGPT recommends your competitor instead of you

A plumber in Newcastle told me his biggest rival shows up every time someone asks ChatGPT for a recommendation, and he never does. They do the same work. He has better reviews. So why does the AI keep handing the lead to the other guy?

It is almost never about who is better. It is about who the model understands and trusts. Here is what is actually going on, and what to do about each cause.

First, see the problem clearly

Before you fix anything, find out what AI actually says about your category. Run a free check and look at who gets named. Then check the Australian AI Visibility Index for your industry and city to see the businesses AI recommends most. You will usually spot the pattern in a minute: the same two or three names, over and over, and a long tail it never mentions.

Now, the reasons.

Reason 1: The AI does not understand what you do

Models recommend businesses they can describe in a sentence. If your website says "premium solutions for the modern home," the model has nothing to work with. Your competitor's site says "ducted air conditioning installation across the Gold Coast," and that is a sentence an assistant can repeat.

The fix: rewrite your opening line and your service pages in plain words. Name the service. Name the suburbs. Drop the adjectives. This is the cheapest, fastest win available, and it helps Google too.

Reason 2: Your competitor is mentioned where it counts

When an assistant names a business, it is repeating what it has read elsewhere. A write-up in a local paper. A genuine review thread. A spot on a directory it trusts. If your rival has three of those and you have none, the model has three reasons to mention them and none to mention you.

The fix: earn a few mentions on sources AI already reads. A customer story in a local publication. A real profile on a reputable directory. You are not chasing volume. You are giving the model evidence.

Reason 3: Your details do not match across the web

Your website says "Smith & Co Plumbing." Your Google profile says "Smith Plumbing Newcastle." A directory has an old phone number. To you these are the same business. To a model, that inconsistency is a reason to hesitate, and a hesitant model picks the business whose story is clean.

The fix: make the name, number and service area identical everywhere. Boring work, real difference.

Reason 4: You are judging it from one assistant

ChatGPT, Claude, Gemini and Perplexity do not agree with each other. You might be invisible on ChatGPT and named on Perplexity. Checking one and assuming the rest match is how businesses misread their position. We track all four every week for exactly this reason.

The fix: look at all four before you draw conclusions. Coverage on one is not coverage everywhere.

A quick way to read the gap

When you run a check, compare your result to the leader in your category like this.

Signal	You	Category leader
Mentioned in how many answers	6 of 80	41 of 80
Named by how many of the four assistants	1	4
Clear service + location on site	Partly	Yes
Mentions on trusted local sources	None	Several

The fixes write themselves once you can see the gap. You are not trying to beat the leader on quality. You are closing the distance on the things the model can actually see.

The order I would do it in

Run a free check and note who gets named instead of you.
Fix your site copy so the service and suburbs are unmistakable. (More on this in our guide to Answer Engine Optimisation.)
Earn two or three mentions on sources AI trusts.
Make your business details identical across the web.
Check again in a few weeks, across all four assistants, and watch the answer change.

None of this is fast magic. It is a handful of unglamorous fixes that compound. The plumber in Newcastle did three of them. Two months later, two of the four assistants started naming him. That is how it goes: not a switch, a shift.

See where you stand with a free AI check, and read how the weekly tracking works if you want the answer to keep changing in your favour.