AI SEO Projects ·Entity Authority
Earned-media entity program
Last reviewed:
- Owner
- Digital PR / comms, with entity-owner coordination
- Metric
- Presence in the outlets most cited by assistants for your category, per product line
- Clock
- Model layer —the two clocks
Only about 5–10% of the references AI cites come from a brand’s own site. In this vertical, that means tech press, gaming media, and creator reviews carry most of the citation weight — and earned coverage is the primary carrier of brand facts into the next training run. This program aims PR at the outlets that actually feed the models, not the ones that feel prestigious.
Why this, mechanically. Earned coverage is slower than owned content and worth more, because it migrates into the model layer where owned edits can’t reach. The neglected half of the work is corrections: models keep repeating stale, high-cited articles, so a correction pathway for the worst offenders is as valuable as new placement.
Deliverable. A ranked list of the outlets assistants cite for your category; a PR program targeting them specifically to refresh what models will “know” at the next training cycle; and correction pathways for the stale high-cited articles that keep resurfacing.
Finish line. The program ships on a cadence tied to model-release timing, and the highest-impact stale citations have a routed correction owner.
This is the slow, durable end of entity work. It pairs with Wikidata stewardship — both feed the model layer, both pay in model-release cycles, not weeks.