Get cited by AI assistants
Make a product page something ChatGPT, Claude and Perplexity can fetch, understand and recommend — measured, not guessed.
- Who it is for
- Founders and marketers who want AI search traffic
- Time
- About 3 hours over a week
- You end up with
- All search and user-fetch crawlers allowed, a correct model reading of your homepage, llms.txt and structured data live, a listing on sources the assistants already read.
- 1
Measure the baseline
Run the crawler access checker. You want the search bots (OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot) and the user fetchers (ChatGPT-User, Claude-User) allowed. Whether you allow training bots (GPTBot, ClaudeBot, CCBot) is a separate decision.
- 2
Fix robots.txt
Generate a robots.txt that names each AI user agent explicitly. A blanket Disallow for GPTBot copied from a 2023 blog post is the most common reason a product is invisible to ChatGPT search.
- 3
Read your site the way a model does
The interpretation checker fetches your homepage without JavaScript and asks a model what the product is. If the answer is vague, the fix is on the page: a text H1 that says what it does, a paragraph for whom, visible pricing.
- 4
Add llms.txt and structured data
llms.txt gives assistants a summary and the canonical links; JSON-LD Organization + SoftwareApplication gives them the facts. Both take ten minutes.
- 5
Be on the pages assistants already cite
Assistants lean on directories, alternatives pages and review sites when asked for recommendations. Your ExitFounder profile, alternatives tags and use cases exist for this; list on the DR-ranked directories too.
- 6
Re-measure and ask
After a week, re-run both checks, then ask the assistants directly with the Ask AI links on your profile. Record the answer; repeat monthly.