SEO for AI search and crawlers without special-markup theatre

Indexable before citable

A page needs a stable URL, successful HTTP response, sensible canonical and crawlable main content. Important statements should not appear only after an interaction or inside an image.

Internal links guide crawlers and people from overview to detail. A sitemap supports discovery but proves neither indexing nor ranking.

Build answers from real substance

Direct openings, clear subheadings and specific checklists improve comprehension. They are useful because people can orient themselves quickly—not because content must be broken into artificial fragments for models.

First-hand expertise appears through decisions, boundaries, failure modes and verifiable sources. Generic summaries add little value regardless of whether a human or model produced them.

Control crawlers deliberately

Public expert pages can be available to Googlebot, Bingbot and OAI-SearchBot. Private administration, personal data and private previews need access control. noindex can exclude public pages from search results but does not make them private.

OAI-SearchBot controls search in ChatGPT; GPTBot concerns model training. These settings are independent. For Google's generative search features, we also check inclusion under Search Console settings. Permission for search is separate from permission for model training.

robots.txt and meta robots serve different purposes. A blocked crawler cannot read a page-level noindex instruction; sensitive material therefore needs access control rather than only a robots rule.

Reject invented AEO shortcuts

Google warns against page batches for every possible query variation and says `llms.txt` neither helps nor harms its rankings. Structured data must describe visible content accurately; more markup does not automatically create more visibility.

We measure Search Console, server requests and genuine referrals separately. A citation in an answer, an impression and a qualified visit are different events.