Why do AI engines care who wrote something?
Grounding is a trust calculation, and authorship is one of its cheapest signals. Two pages make the same medical, legal or technical claim; one is by a named professional with a consistent footprint — credentials, other publications, professional profiles, mentions — and one is by “Admin”. The engine does not need to judge the medicine; it judges the provenance. This is the AI-era continuation of EEAT: experience, expertise, authoritativeness and trust, evaluated at machine speed. Anonymous content is not penalized explicitly — it simply loses ties against attributable content, constantly, at scale.
What makes an author a machine-readable entity?
- A canonical author page. Name, role, credentials, photo, topics, and links to external profiles — the hub that ties the identity together.
- Consistent naming. The same name form everywhere: site, LinkedIn, conference bios, guest articles. Variant spellings fragment the entity.
- External corroboration. Professional profiles, speaking pages, podcast appearances, citations of their work — third-party proof the person and expertise exist.
- Structured data. Article markup with author properties pointing to the author page; Person markup on the page itself. See structured data for AI.
- Topical consistency. An author who writes coherently within a domain builds topical authority engines can detect; a byline attached to everything builds none.
How do author entities interact with organization entities?
Authors ladder up to the brand. An article’s credibility combines the author entity, the publishing organization’s entity, and the sources cited. Strong author entities lift the whole domain’s grounding confidence; conversely, a strong brand cannot fully rescue content with no attributable human behind it. The organization side — Organization markup, Wikidata, knowledge panels — is covered in entities and the knowledge graph and Wikipedia and Wikidata.
How do you build author entities in practice?
- Inventory your real experts: who actually knows the material you publish?
- Create or upgrade author pages with credentials and external links.
- Retrofit bylines: replace “Team” and “Admin” with named humans on your highest-value content first.
- Build external footprints deliberately: guest articles, talks, community participation. See community seeding.
- Monitor: include author-attributed content performance in your monthly citation measurement.
Common pitfalls
The worst practice is fabricated authors — AI-generated personas with stock photos. Grounding cross-checks fail: no external footprint, no corroboration, and discovery converts the whole domain’s provenance into a liability. Second, byline sprawl: dozens of one-off names with no entity building behind any of them. Concentrate authority in real people. Third, letting departing employees take the entity with them: keep author pages accurate about current roles, and reattribute gracefully. Finally, remember authors matter most where trust matters most — YMYL topics, technical depth, original research — prioritize entity building there, as detailed in EEAT for AI.
Scale the program to your reality. A solo founder does not need ten author entities — one strong one beats ten hollow ones. A large publisher should prioritize entity building for authors covering trust-sensitive topics first. In every case, the test is external resolvability: ask an AI engine who your author is. If the answer is accurate and cites real credentials, the entity is built; if the engine shrugs, the footprint work is not done.