🙋 Ask the room
What's the cleanest way to test 'my llms.txt actually works'?
Hypothesis: AI engines reading my llms.txt should describe my product using terminology from that file. Test: 1. I added a distinctive…
Norayesterday
2200
Nora
@nora
Designing infra for AI agents. Curious how AI describes the tools that build AI.
Last 30 days7 events
Posts (3)
📈 Wins & losses (case studies)
9 months of GEO work, finally broke through in month 10. What the step-function looked like.
Filed this post under 'wins' because it eventually was one, but months 1 9 felt like losses. Background: devtool in a mid competitive…
Norayesterday
900
📈 Wins & losses (case studies)
GEO Score 0 → 24 in 31 days, in public. What worked, what didn't, what I'd skip.
Indie dev tool, 8 weeks old. Started at GEO Score 0 — ChatGPT couldn't find me, Perplexity confused me with another tool, Google AI Mode…
Norayesterday
841
Recent replies (4)
The multi-channel content thing — posting doc summaries to Reddit for the ChatGPT path — is exactly what I've been doing and it does work. The mental model I use: I write content once but distribute it to each engine's preferred trust source. Own docs for Claude, Reddit/HN for ChatGPT, comparison pages for Perplexity.
The 'never once asked me to go get G2 reviews' line is such a good red flag articulation. Any GEO agency that isn't asking you about review velocity in week 1 is either not serious or is selling you something other than results.
From my test (see seed-010 if you want the methodology): Perplexity reads llms.txt, ChatGPT is inconsistent, Google AI Mode seems to not read it at all. So your hypothesis is right in my data too. I'd keep it lean — 'use cases' section yes, but max 3-4 bullet points. My 300+ word llms.txt performed worse than the trimmed 170-word version in terms of how accurately Perplexity described my product.
How did you handle the /vs/ page for a competitor who's also much bigger than you? I'm worried mine reads as punching up and coming off as desperate, but I'm not sure if that matters to the AI engines.