How to Get Your SaaS Product Recommended by ChatGPT and AI Search

SaaS products get recommended by ChatGPT and other AI assistants mainly because of what other websites say about them: comparison articles, "best tools" lists, review platforms and community discussions. Their own websites matter too, especially clear, specific product and comparison pages. No setting, file or schema markup gets a product recommended on its own.

Key takeaways

• AI assistants build software shortlists largely from third-party pages: "best X" lists, comparisons, reviews and forums.
• In one widely cited study, "best X" list pages were the most common type of page cited in AI answers about software and services.
• Your own site still matters: specific, quotable product, pricing and comparison pages are what the assistant can confirm facts from.
• llms.txt files and schema markup have no demonstrated effect on AI recommendations.
• Track it by asking the assistants regularly and by adding "AI assistant" to your "How did you hear about us?" field.

How AI assistants build a software shortlist

When someone asks "What's the best [category] tool for a 20-person agency?", assistants with web access typically run several searches behind the scenes, read a set of results, and summarise. Google describes its AI features as using a "query fan-out" technique: issuing several related searches to answer one question. Which products appear depends heavily on which pages those searches return and what those pages say.

For software questions, those pages are disproportionately:

• "Best [category] tools" listicles
• Head-to-head comparisons and "alternatives" pages
• Review platforms in your category
• Community threads (Reddit, specialist forums) where users compare tools

Ahrefs analysed the sources AI assistants cited for "best X" prompts across software, products and agencies, and found "best X" list pages made up the largest share of cited page types: about 43.8% in that sample (Ahrefs). It's one study from one vendor, but it matches what you see when you ask the assistants yourself.

What you can actually do

1. Get included in genuine comparison content

Find the "best [category]" and "[competitor] alternatives" articles that already rank in Google for your category. These are what the assistants read. Then:

• Contact the authors with accurate, current information about your product
• Offer a trial or demo so they can evaluate it properly
• Correct outdated information about you where it exists

Don't pay for fake reviews or placements disguised as independent recommendations. Undisclosed paid endorsements can break advertising rules, and they're fragile when platforms crack down.

2. Build your own honest comparison pages

"[Your product] vs [competitor]" and "[competitor] alternatives" pages on your own site rank for high-intent searches and give assistants something to cross-check. Be fair: state where the competitor is stronger. A page that only praises you reads as advertising to humans and machines alike.

3. Be present on review platforms buyers use

If your category has established review sites, a profile with genuine reviews from real customers matters. Ask happy customers directly, without incentives that platforms prohibit.

4. Make your own pages quotable

Assistants often work from short excerpts. One 2026 technical study of ChatGPT's search found the model mostly sees a page's title, URL and a short snippet, and fully opens only a small fraction of pages (RESONEO). So:

• Put a one-sentence description of what the product does, and for whom, near the top of key pages
• State specifics: pricing, integrations, limits, compliance, supported regions
• Use headings that say plainly what the section is about

Want to know what ChatGPT says about your product right now? We'll run the prompts your buyers use on a free strategy call and show you which sources the answers rely on. Book it.

Related: Want to measure where you stand before changing anything? Our AI visibility audit guide shows how to test ChatGPT, Perplexity and Google AI with a repeatable prompt set.

5. Earn mentions in places people discuss your category

Genuine participation in communities, podcast appearances, guest articles and PR all create the third-party mentions assistants draw on. Ahrefs found branded web mentions correlated more strongly with AI Overview visibility than backlinks did, across 75,000 brands (Ahrefs). That's correlation, among established brands, but it points the same way as the rest of the evidence.

6. Make sure AI crawlers can read your site

If your site or CDN blocks OpenAI's search crawler (OAI-SearchBot), PerplexityBot or similar, those assistants can't read your pages properly. OpenAI states that sites which don't block OAI-SearchBot can appear in ChatGPT search (OpenAI). Check your robots.txt and firewall bot settings.

What doesn't work

• llms.txt files. Google has said it doesn't use them, and there's no evidence the major assistants rely on them to choose recommendations.
• Schema markup as an AI lever. A 2026 study of 1,885 pages that added structured data found no clear improvement in AI citations. Schema has other uses; this isn't one.
• "Guaranteed placement in ChatGPT." Nobody can promise this. Answers vary by prompt, user, location and over time.
• Mass-produced AI content. Publishing many thin pages at scale risks Google's spam policies, which reduces the visibility assistants draw on.

How to measure it

1. Write 10–20 prompts your buyers would realistically ask ("best [category] for [use case]", "[competitor] alternative", "[category] for [industry]").
2. Run them monthly in ChatGPT, Perplexity, Google AI Mode and Gemini. Repeat each a few times, because answers vary.
3. Record whether you're mentioned, how you're described, and which sources are cited.
4. Add "AI assistant (ChatGPT, Perplexity, etc.)" to your "How did you hear about us?" field.

The cited sources are the most useful part: they show which pages to get into or improve.

Frequently asked questions

How do I get my software recommended by ChatGPT?

Get included in genuine "best [category]" lists, comparisons and review platforms that rank for your category; publish fair comparison pages on your own site; earn mentions in communities and media; and make sure AI crawlers can access your site. No single action guarantees a recommendation.

Where does ChatGPT get software recommendations from?

When it searches the web, ChatGPT draws on the pages its searches return, frequently "best tools" lists, comparison articles, review sites and community discussions, as well as vendor websites for specific facts.

Does llms.txt help get my SaaS mentioned in AI answers?

There's no evidence it does. Google has said it doesn't use llms.txt, and large-scale analyses have found no relationship between having the file and being cited.

Can an agency guarantee my product appears in ChatGPT?

No. AI assistants' answers vary by prompt, user and time, and no one controls them. An agency can improve the inputs (third-party mentions, comparisons, crawlability, clear content) and measure the results honestly.

How long does it take to appear in AI recommendations?

It depends on how quickly third-party coverage changes. Fixing crawler access and your own pages can have an effect fairly quickly, because assistants read pages live; earning inclusion in comparison articles and reviews usually takes months.

---

Want your product in the answers buyers get from AI? Our AI search optimisation work focuses on what has evidence behind it, measured honestly.

Book a free strategy call · GEO services · AI SEO services · The GEO guide