Why ChatGPT Doesn't Recommend Your SaaS
Why doesn't ChatGPT recommend my product?
Usually not because your product is worse. It's because the model either doesn't have you in its picture of your category, or it can't verify enough about you to say your name with confidence. Models are cautious by design: when they're unsure, they fall back on the option they can describe precisely — which is almost always your better-documented competitor. It helps to picture what happens when someone types "best [your category] tool for [use case]." The model isn't running a live search of the whole web and ranking products by quality. It's assembling an answer from what it has absorbed about your category and, in some products, a handful of pages it retrieves in the moment. If you're thinly or inconsistently represented in that material, you simply don't make the shortlist — regardless of how good the product is.
What does ChatGPT "check" before recommending a tool?
There's no literal checklist inside the model, but in practice a recommendation depends on a few things you can influence. Presence in the sources models read. Your own site, yes — but also directories, review sites, forums, and communities. Models weight what others say about you heavily, because third-party corroboration is harder to fake than your own homepage. Extractable positioning. Can the model lift a clear sentence that says what you do and who it's for? "The lightweight CRM for solo real-estate agents" is extractable. A hero section that just says "Close more. Stress less." is not — it tells the model nothing it can repeat. Consistency across sources. If your homepage, your G2 listing, and your Twitter bio describe you three different ways, the model can't form a stable picture, so it hedges by leaving you out. Category-word match. If buyers ask for an "AI visibility tool" and you describe yourself as an "agentic discoverability intelligence platform," you've made yourself hard to match to the actual question.
The most common reasons you're invisible
In practice, the same handful of problems come up again and again. Your value proposition is trapped in images, video, or JavaScript the crawler never renders. If your core message only exists inside a graphic or loads client-side, to a text-based crawler the page is nearly blank. You invented a category no one searches for. Clever naming that doesn't map to how buyers actually ask is a self-inflicted wound. No third-party corroboration. You're not on the directories, review sites, or communities models treat as evidence, so there's nothing to confirm your claims. Inconsistent descriptions across the web, so the model can't settle on a single stable picture of you. You're blocking the crawlers — sometimes accidentally, via an over-aggressive robots.txt or bot protection that shuts out GPTBot and friends.
How do I get ChatGPT to recommend me?
Make yourself easy to describe and easy to verify, then measure it. Start on your own page: state, in plain text near the top, exactly what you do and who it's for, using the words your buyers use. Add structured data so the model can parse your product, pricing, and FAQs. Publish an llms.txt that maps your key pages. Confirm you're not blocking AI crawlers. Then work outside your own walls, because that's where the trust lives: get listed — with the same one-line description everywhere — on the directories, review platforms, and communities relevant to your category. Consistency is the multiplier; the identical description repeated across many sources is what lets a model say your name with confidence. Finally, measure. You can't improve what you can't see, and "are we being recommended?" is answerable. That's exactly what an Agentic Score is for: it tells you which of the reasons above is holding you back, so you fix the real problem instead of guessing.
Frequently Asked Questions
- Why does ChatGPT recommend my competitor but not me?
- Almost always because the model can describe your competitor more confidently — they're more consistently documented across the sources it reads. It's a legibility gap, not necessarily a quality gap, and it's fixable.
- Does ChatGPT actually browse my website in real time?
- It depends on the mode. Base models answer largely from what they absorbed in training; browsing or retrieval features can pull live pages. Optimizing for both means being well-represented in training-era sources and easy to parse when fetched live.
- How long until ChatGPT starts recommending me after I make changes?
- Structural fixes (schema, llms.txt, crawler access) can register whenever models next crawl or retrieve your pages. Reputation and corroboration compound more slowly, over weeks and months.
- Will adding an llms.txt make ChatGPT recommend me?
- On its own, no single file flips a switch. llms.txt removes ambiguity about your site, and combined with consistent descriptions and third-party presence it meaningfully improves your odds of being described correctly.
- Is this the same as ranking #1 on Google?
- No. You can rank well on Google and still be skipped by ChatGPT, because they use different signals. Google rewards pages; AI models reward tools they can describe and verify.