5 Reasons ChatGPT Isn't Recommending Your Brand (and How to Fix Each)

You ran the test. You asked ChatGPT to recommend the best options in your category, and your brand wasn’t in the answer. Or it was, buried at the bottom, described with a shrug.

Annoying. Also fixable. In our experience it’s almost never bad luck and almost always one of five specific problems. Here they are, ranked from most common to least, with the fix for each.

1. The AI literally cannot read you

This is the most common reason, and the most embarrassing, because it’s invisible until someone checks.

AI engines send their own crawlers: GPTBot, ClaudeBot, PerplexityBot, Google-Extended. Plenty of sites block them by accident, through an overzealous robots.txt, a firewall rule, or a security plugin nobody remembers installing. If the crawler can’t get in, you are not “ranked low.” You do not exist.

The fix: Confirm your robots.txt allows the AI crawlers. Make sure your important pages render without JavaScript gymnastics the crawler can’t follow. Add an llms.txt that points the models at your best pages. This is the floor. Until it’s solid, nothing else you do will show up.

2. You’re a brochure, and the AI knows it

AI does not trust what you say about yourself. It trusts what the rest of the web says about you. A brand that only appears in its own marketing reads as unproven. A brand mentioned across reviews, roundups, and real conversations reads as legitimate.

If the only place your name shows up is your own domain, the model has nothing to corroborate, so it plays it safe and recommends someone with more outside validation.

The fix: Get cited somewhere other than your own site. The “best of” lists in your category. Review platforms. The communities where your buyers actually hang out. You’re not gaming anything. You’re giving the model the second opinion it needs before it will vouch for you.

3. Your content can’t be quoted cleanly

AI doesn’t lift whole pages. It lifts the single most relevant passage it can quote with confidence. If your value is buried in clever prose, long intros, and vague claims, there’s nothing for the model to grab.

The brands that get pulled into answers tend to write in a way that’s easy to extract: a clear question, a direct answer right under it, facts in tables, specifics instead of adjectives.

The fix: Restructure your key pages to answer questions plainly and immediately. Put comparisons in tables. Lead each section with the answer, then support it. Write for the passage, not the paragraph. If a model can lift one clean sentence that makes your case, you’re quotable.

4. The AI believes something outdated or wrong

Sometimes you’re in the answer, but the answer is off. Old pricing. A category you left years ago. A feature you killed. Or a quiet negative lean anchored to one stale review the model keeps surfacing.

This happens because the model learned something once and hasn’t been corrected. It’s repeating an old version of you.

The fix: Find the source. Figure out whether the belief is coming from a specific page it’s retrieving live (faster to fix) or baked into its training (slower, needs a sustained campaign of fresh, consistent signals). Then correct the record at the source so the next answer reads differently. This is detective work, and getting the diagnosis right is most of the battle.

5. You’re missing from the sources that engine trusts

Here’s the one that catches people who think they’ve done everything right. You might be well represented for ChatGPT and invisible in Perplexity, because the two engines lean on different sources.

Each engine has its own taste in what it cites. Win the places ChatGPT trusts and you can still be absent from the answer Gemini gives, in the same category, to the same buyer. People who “optimized for AI” and only checked one engine fall straight into this.

The fix: Treat each engine as its own project. Map which sources each one actually pulls from in your category, then earn presence on the ones you’re missing. It’s more work than a single playbook, and it’s why coverage across all the major engines is the real standard, not a nice-to-have.

It’s usually two or three at once

The frustrating truth: most brands aren’t tripped up by one of these. They’re tripped up by a few, stacked. You fix the crawler problem and realize you also have no third-party proof. You build the proof and realize your pages aren’t quotable.

That’s normal. Work them in order, top to bottom, because the earlier ones gate the later ones. A perfectly quotable page the crawler can’t read still gets you nothing.

Want the diagnosis done for you? Get a free AI visibility audit and we’ll tell you exactly which of these five are holding you back, across every major engine, with a plan to fix them. Or book a call and we’ll walk through your results together.