Why AI keeps recommending your competitor — and how to change it
Ask ChatGPT, Gemini or Perplexity for the best option in your category and you get a tidy little shortlist — three or four names, a sentence of reasoning each. If your competitor is on it and you're not, it isn't because their product is better. It's because the model found them in the pages it read, and didn't find you. The recommendation is assembled, not judged — which means it's something you can actually influence.
The decision happens without you in the room
A buyer used to open ten tabs and make up their own mind. Now a growing share of them ask an assistant — "what's the best CRM for a small agency?", "who should I use for home insurance in Ireland?" — and take the shortlist it hands back. That shortlist is the new shelf. You're either on it or you're invisible, and unlike a Google results page there's no page two to scroll to. Three names, maybe four, and the conversation moves on.
It picked them because it read about them
Modern answer engines don't hold an opinion about your market. When asked to recommend, they run live web searches and paraphrase what comes back — so the shortlist is really a summary of whichever pages the model happened to retrieve. Your competitor shows up because they're named in the comparison articles, the "best of" listicles, the Reddit threads and the review sites the model pulled. If those pages don't mention you, no amount of being genuinely good gets you into the answer. The uncomfortable version: the model is recommending the sources' favourite, not the market's best.
Four sources do most of the deciding
Different engines lean on different mixes — ChatGPT weights Wikipedia and mainstream media heavily; Perplexity leans hard on Reddit and fresh, structured pages; Google's AI Overviews track the same E-E-A-T and schema signals as classic Search. But the raw material is remarkably consistent. Get named in these and you get named in the answer.
| Source type | Why the model trusts it | Your move |
|---|---|---|
| Comparison & "best of" articles | Pre-structured as a ranked list it can lift wholesale | Get included — pitch reviewers, earn round-up spots |
| Community threads (Reddit, forums) | Reads as unpaid, real-user experience | Be genuinely present and recommended, not astroturfed |
| Review platforms | Aggregate sentiment and recency it can cite | Ask happy customers; keep the profile current |
| Wikipedia & reference | Treated as neutral ground truth | Earn the coverage that justifies an entry |
Fix the sources, and the answer fixes itself
You can't argue with the model, and you can't optimise your own site into an answer that never reads it. The lever is the third-party record. Concretely:
Most businesses have never actually looked
Here's the part that should bother you: almost nobody checks. They assume that ranking on Google means showing up in AI answers, and the two have quietly come apart. The only way to know whether an assistant recommends you, ignores you, or — worse — recommends a competitor while getting your details wrong, is to ask it the questions your customers ask, across the engines they use, and watch what comes back over time.
Image (brand-treated from licensed original): High-street shop parade by Acabashi, CC BY-SA 4.0, via Wikimedia Commons. Treated versions of the CC BY-SA original are shared under the same licence.
