Why does AI recommend my competitors and not me?
Almost always one of four reasons: the web describes your rivals as places to buy and not you; your site can't be read or cited by AI crawlers; your name is ambiguous so mentions don't attach to you; or you're being out-mentioned in the community sources AI trusts. Each has a different fix — diagnose before spending.
Updated 22 July 2026
It is the question that starts most AI-visibility projects, usually asked with some heat: a customer mentions that ChatGPT recommended a rival, someone checks, and there you are — absent. Before spending anything, work out which absence you have, because the four causes have four different fixes and effort on the wrong one is wasted.
Cause 1: the web doesn't describe you as an answer
Assistants learn who belongs in an answer mostly from third-party writing: roundups, reviews, trade press, forums. If "best [your category] UK" articles name three rivals and not you, the assistants are not being unfair — they are accurately reflecting a web that has not written you into the category.
Test: search for the roundups and listicles in your category. Count who appears. Fix: earn presence in them — outreach, genuinely distinctive offerings, the slow editorial work. This is the biggest lever and the slowest, and no amount of on-site work substitutes for it.
Cause 2: your site can't be read or cited
If your pages block AI crawlers, need JavaScript to render, or bury the answer to "what is this business?", then even when an assistant searches mid-answer it cannot lift you into the result. Your rivals' readable pages get cited; your unreadable ones do not exist.
Test: run a free scan — crawlability, rendering and structured data are checkable in minutes. Fix: this is the fast half. Server-rendered pages, AI crawlers allowed, clean structured data, one clear self-contained claim near the top of each key page.
Cause 3: your mentions don't attach to you
A brand with a generic or ambiguous name can be mentioned and get no credit — the words appear, but nothing ties them to your business rather than the phrase itself. We measure this directly: our own index carries an attribution filter precisely because we suffered the problem ourselves.
Test: ask an assistant about your brand by name plus location or category. If it describes a different business — or a generic concept — you have an entity problem. Fix: consistent name, address and description everywhere; Organization structured data with sameAs links to your real profiles; the disambiguating word in your public name where feasible.
Cause 4: you're losing the community conversation
The sources assistants cite most are communities and enthusiast sites, and if the threads recommend rivals, the answers will too.
Test: search Reddit and the enthusiast forums for your category. Read who gets recommended and — more usefully — why. Fix: participate honestly, deserve mentions, and treat the stated reasons rivals win threads as free market research.
Diagnose, then spend
The order matters: cause 2 is checkable today and cheap to fix; causes 1 and 4 are the long game that most brands should start immediately precisely because it is slow; cause 3 is rare but severe when present. A measured reading — which engines name you, at what rate, against which field — tells you which absence is actually yours before you buy a fix for someone else's.