AI assistants recommend products they can read about in several independent places that agree with each other. Let their crawlers in, state plainly on your own site what the product does and who it's for, and get described the same way on directories, launch platforms, comparison articles and forums. Nobody can guarantee a mention, and anyone selling one is guessing.
How assistants pick what to recommend
An assistant answers from two sources. One is what it learned in training, which is months old and can't include a product launched last week. The other is a live web search it runs while answering, then reads the top results. A new product gets recommended almost entirely through the second route.
So the practical question is: when the assistant searches for "best tool for X", do the pages it finds mention you, and do they say the same thing about you? That's most of the work.
Let the crawlers in
- Check robots.txt. Some site templates and security settings block AI crawlers by default. If you want to be cited, allow them.
- Serve real HTML. Many AI crawlers don't run JavaScript, so content that only appears after scripts load is invisible to them.
- Submit your site to Bing Webmaster Tools as well as Google Search Console. Assistants lean on more than one search index.
- Add an llms.txt file at the root of your site: a short plain-text summary of what the product is, with links to the key pages. It's a proposed convention, not a standard every assistant reads, but it's quick to add.
Say plainly what you are
Assistants quote sentences. Give them one worth quoting: what the product is, who it's for and what it costs, near the top of the home page, in plain words. "Invoicing software for freelance designers, free for up to five clients" can be lifted straight into an answer. "Reimagine your workflow" can't.
- Put pricing on a public page. Assistants are asked for free and cheap options constantly, and skip tools whose price they can't find.
- Write a page for each main use case and each comparison people ask about, such as you against the well-known tool.
- Add an FAQ in question-and-answer form. It matches how people phrase requests to an assistant.
Be described the same way in several places
One page saying you're good is a claim. Five independent pages saying the same thing is evidence, and that's what an assistant weighs. Use the same name, the same one-line description and the same category everywhere you're listed.
- Directories and launch platforms. They're crawled often and structured the same way for every product. Start with our list of free startup directories, and the AI tool directories if AI does the main job in your product.
- "Best tools for X" articles. Assistants read these heavily for recommendation questions. Find the ones in your category and ask the author to consider you.
- Forums. Reddit threads and community answers are cited often. See where to post your startup on Reddit.
- Reviews and comparison sites where real users describe what the product does.
How to check whether it's working
- Ask several assistants the questions your customers would ask, without naming your product, and note which products and sources come back.
- Look at the sources they cite. Those pages are your target list: get listed or mentioned there.
- Watch your analytics for visits referred by assistant domains, and ask new signups how they found you.
- Repeat monthly. Answers shift as indexes update.
Get a page assistants can read
IndieLaunch product pages are open to AI crawlers and state what your product does, who made it and what it costs. Launching is free.
Frequently asked questions
- Can I pay to be recommended by ChatGPT?
- No. You can't buy a place in an assistant's answer. Be wary of services that promise one; what they can actually do is the listing and content work described here.
- Does llms.txt help?
- It may. It's a proposed convention that some tools read and others ignore. It takes a few minutes to write, so it's worth adding, but clear pages and outside mentions matter more.
- How long until a new product shows up in AI answers?
- Through live search, as soon as the pages that mention you are indexed, which can be days or weeks. Through the model's own training, not until a later model is trained, which you can't control.