AI recommends my competitor instead of me
Ask an engine for the best option in town and a few names come back. If theirs is on the list and yours is not, there is usually a specific and fixable reason. Here is how to find out which one, rather than guessing.
Last reviewed
Start by asking why, literally
The single most useful thing you can do takes two minutes. Ask the engine for the recommendation, then ask it to justify itself:
- “Which are the best [category] in [town], and why?”
- “Why did you recommend those and not [your business name]?”
- “What sources did you use?”
The reply is not gospel — a model explaining its own reasoning is producing a plausible story, not a log. But it is remarkably often concretely useful, because it surfaces which pages it read. Those pages are the actual battleground.
The five reasons, in the order they usually apply
1. It does not know enough about you to risk naming you
This is the most common by a distance, and it does not feel like a ranking problem — it feels like invisibility. A model asked to recommend is making a claim it can defend. If your evidence is thin — sparse Business Profile, few recent reviews, a website that says little in extractable text — you are a risky recommendation, and it names someone safer.
If your business does not come back with a solid description when you ask about it by name, this is your problem, and why your business is not showing up in AI search is the guide for it. Nothing below matters until this is fixed.
2. Their reviews say something yours do not
Look at the actual text of their recent reviews next to yours. The question was probably specific — best for families, best for a quick lunch, best for emergency callouts — and a model can only name you for a specific thing if somebody wrote that thing about you.
You cannot write those reviews. You can ask customers, when they are happy, what they came in for — people tend to answer that question in the review they leave. Over a couple of months that changes the vocabulary attached to your name.
3. They are in a list you are not in
Search “best [category] in [town]” yourself and read the roundups on page one. Local newspaper lists, blog roundups, “top 10” posts, chamber of commerce pages, event write-ups. These are exactly the sources a model leans on for a “best of” question, because they are literally that question already answered.
Being added to two or three of them is the highest-leverage thing on this page, and it is mostly a matter of asking. Most of these lists are maintained by someone reachable.
4. Their category is more precise
If they are listed as “Neapolitan pizza restaurant” and you are “Restaurant”, they will win pizza questions and you will compete with everything that serves food. Primary category does an enormous amount of quiet work. Check what they have chosen — it is visible on their profile — and consider whether yours is as specific as it honestly can be.
5. Something about you contradicts itself
Two different phone numbers across directories, hours that disagree between your site and your profile, an old address still live somewhere. Contradictions do not average out; they make you the ambiguous option. See fixing wrong information about your business for how to trace each one to its source.
Write down who gets named
Keep a dated list of the businesses that come back, every time you ask. It is worth more than it sounds.
The list is your competitive set as the engine understands it, which is frequently not the one you assume — it often includes a business you do not think of as a rival and omits the one you watch closely. And when the list changes, that is the earliest signal you will get that something moved, in either direction.
What not to do
- Do not report a competitor out of frustration. Report a genuine rule breach — a keyword-stuffed name, obviously fake reviews — and nothing else. Frivolous reports go nowhere and waste your afternoon.
- Do not buy reviews to catch up. Detectable, removable, and a filtered profile helps nobody.
- Do not argue with the chatbot. Nothing typed into a chat window teaches the model anything about your business.
Then watch whether it changes
Everything above takes weeks to show up, and the only proof is the answer changing. Pro plans on AI Business Monitor track exactly this: who gets named when the engine is asked for the best in your town, week by week, so you can see a competitor arrive or drop off instead of finding out months later.
Common questions
Is my competitor paying for this?
Almost certainly not. There is no paid placement inside an AI recommendation today. If they are being named and you are not, something in the evidence favours them — and evidence is something you can change.
They have fewer reviews than me. Why are they winning?
Count is only one input, and often not the strongest. Recency matters, and so does what the reviews actually say. Forty recent reviews describing specific things beat two hundred old ones saying "great service", because the specific ones can be quoted in answer to a specific question.
Can I get a competitor removed from the results?
Not unless they are breaking a platform rule — a keyword-stuffed business name or fake reviews can legitimately be reported to Google. Otherwise the answer is to become the better-evidenced option, not to try to remove them.
Does asking the question repeatedly change the answer?
No. Answers vary between accounts and sessions naturally, but nothing you do inside a chat window trains the model about your business. Ask twice to see the variation; do not expect to influence it.
Check it every week automatically
Doing this by hand once tells you about today. Things change: hours get edited, reviews land, a competitor opens. We check weekly and email you when something moves, with the source behind every finding.