Start here if ChatGPT, Claude, Gemini or Perplexity recommend your competitors and not you. These guides cover how assistants choose, how to measure where you stand, and which fixes have evidence. Every claim links to its source. Every guide says where the evidence is thin.
The guides
- How AI decides what to recommend: it searches first, reads what others say about you, and breaks ties for the brand it knows. The evidence, in order.
- How to check if ChatGPT recommends you: our free check, done by hand. Five question shapes, three runs each, five assistants, and what to write down.
- Why AI doesn’t recommend your product: five causes in the order they usually happen, each with a test you can run today.
- What works and what doesn’t: third-party proof and reviews work. llms.txt alone, schema alone, keyword stuffing and directory blasts do not. Studies for each.
- Why a glossary page helps AI find you: one definitions page, about 30 to 120 search impressions a day in two weeks. What I built, and the honest link to AI.
- Watching the competitors AI names instead of you: a $7.80 a month agent that checks pages every morning, and what it caught in two weeks.
- AI-drafted support replies, done honestly: 14 hours to 12 minutes, and the rules that kept it honest. Support is where the reviews AI reads come from.
- From side project to a product AI can name: five sessions turned an app for my mom into nine products. The sixth step makes one an assistant can recommend.
Coming later
Getting into the “best X” lists AI cites. Getting recommended by Claude, and by ChatGPT, specifically. Making your own site readable by AI. We add a guide when we have something sourced to say, not before. New guides go in the changelog.
How these guides are written
Dasha writes them, with AI help on first drafts. Every number is checked against the original study before it goes up. Vendor studies (from companies that sell AI visibility tools) are labelled as vendor studies. Found a mistake? Email hello@heardofyou.io. We fix it and log the fix.