At least one valid JSON-LD block is present.
Agents prefer JSON-LD because it is typed and machine-parseable. Microdata and RDFa get ignored. Sites without Organization or WebSite schema lose brand recognition in agent answers, because agents have to infer the publisher from page text instead of citing a structured source.
Underlying specification: https://schema.org
- Run the AgentSpeed structured-data validator to see which JSON-LD blocks (if any) your page already exposes.
- Add an Organization block in your <head> on every page, with name, url, logo, and sameAs links to your social profiles.
- Add a WebSite block with potentialAction → SearchAction so agents know how to query your site.
- For product / article / FAQ / breadcrumb pages, add the matching schema.org type.
- Validate with the validator after deploy; copyable templates are generated for missing schemas.
The snippet below uses your-domain.example as a placeholder. Replace it with your actual host before deploying. For a fix tailored to your live site, run a scan and open the per-domain remediation page.
Paste this into Claude Code, Cursor, or any assistant with your repository open. It carries the steps, how to verify them, and the shortcuts to avoid. It is generic, though: run a scan to get one written from what your pages actually do.
Do I need both schema.org and Open Graph?
Yes. They are read by different consumers. JSON-LD / schema.org is read by AI agents and Google for rich results. Open Graph / Twitter Cards are read by social platforms for link previews. Neither replaces the other.
Does invalid JSON-LD hurt SEO?
Indirectly. Google ignores invalid blocks for rich results. The bigger cost is lost agent visibility: partial Organization or WebSite blocks force agents to fall back to text inference for the publisher name.
Run a scan against your domain to see whether structured_data.jsonld_present currently passes, warns, or fails. Free, no signup; the score page is permanent and citable.