How AI Crawlers Work
AI systems fetch the web for training, for live retrieval, and for answering a question you asked a moment ago. Those are three different crawlers with three different consequences.
Intermediate8 min readDaksh BathlaA proposed convention: a Markdown file at a site's root pointing AI systems at a clean summary of its most useful content. It is not a standard and grants no permissions.
A file listing your twenty most useful pages with one-line descriptions, so a model doesn't have to reconstruct what your site offers from navigation and marketing copy.
Adoption is partial and voluntary, so treat it as cheap and low-risk rather than decisive. The real benefit is that writing it forces you to state which pages actually matter.
AI systems fetch the web for training, for live retrieval, and for answering a question you asked a moment ago. Those are three different crawlers with three different consequences.
Intermediate8 min readDaksh BathlaGenerative engine optimisation: being findable when the answer is written by an AI system instead of listed as ten blue links.
Intermediate7 min readDaksh Bathla