R6 · ANSWER-ENGINE OPTIMISATIONPROTOTYPEMar–Apr 2026

Making businesses readable to AI answer engines

THE QUESTION

Can a store's product catalogue be turned into a dense, factual document that AI crawlers and answer engines can read easily?


Founder, Bodham LabsUpdated

What did we build?

  • NeoIndex: a Wix app prototype. It reads a store's products (up to 50) and uses Gemini 2.0 Flash to write a dense, factual Markdown document for AI crawlers.
  • A read endpoint, and an OpenAI-compatible chat endpoint.
  • Bring-your-own-key: the merchant's own API key is stored encrypted (AES-GCM).
  • Its one-line goal: "Convert product catalogs into dense, factual Markdown documents for AI crawlers."

What does the evidence show?

Values exactly as recorded. The source column says where each one was recorded.
WHATVALUESOURCE
Products read per storeup to 50app source
Commits3, Mar–Apr 2026repository history
DeploymentNever deployed or listedrepository history

Where has it led us?

  1. The prototype was never deployed or listed, so we have no data on whether it changed what answer engines say.

  2. We do apply the same ideas live on bodhamlabs.com: an llms.txt file, JSON-LD structured data, and robots rules for AI crawlers.

What is still open?

  • Does a dense Markdown summary change what an AI answer engine says about a business?
  • How would we measure that fairly, given that answers change from one day to the next?

NEXT STEP · NOT DONE YET

What would we do next?

Track how answer engines describe bodhamlabs.com over time, before offering the same approach to anyone else.

Have a problem that needs this kind of research?

Tell us about it