Technical discoverability
Everything in this pillar is a precondition rather than a strategy. A technically sound site earns no citations by itself. A technically broken one is excluded before the contest begins, and the owner rarely finds out. This is the plumbing: unglamorous, checkable, and most of it can run without you.
- Retrievability is binary and silent. A blocked crawler, a firewall rule, or content that only exists after JavaScript runs: each excludes you without an error message you will ever see. Required
- One canonical URL per thing. Split versions split your standing. Official recommendation
- Text is the retrievable form. Facts locked in images or interactive widgets are facts a retrieval system may never see. Official recommendation
- This layer belongs to your platform. None of it needs the owner's judgement, so hosting platforms handle it most fully, bernard included. Context-dependent
In this pillar
- Can AI reach your site at all?live
- Static HTML & JavaScript renderinglive
- robots.txt in practicelive
- XML sitemaps that stay truelive
- Canonical URLslive
- Redirects & status codeslive
- Accessibility & agent-friendly siteslive
- Duplicates · Internal links · Performanceplanned
What bernard does about this pillar
This is bernard's home ground: hosted sites get static HTML, managed robots and sitemaps, canonical URLs and structured-data validation as standard. verified
Standing audits for redirect chains and status-code drift are partial today; the gap is tracked openly. partial
- Contradicted “Create an llms.txt file to rank in Google's AI results”
- Contradicted “Add special AI schema to your pages”
Each verdict carries its primary source and the date we last checked it. See the full ledger.