Search engines and AI assistants don’t read your website to decide who you are — they read the entity graph. Wikidata is the open spine of that graph. So across every personal brand we publish, we run one disciplined question on a schedule: does this person have a Wikidata item, should they, and is it accurate and linked? This is the system that answers it — audited weekly, not once.
The one-line version: We audit the whole fleet for Wikidata coverage every week, link and enhance the people who are genuinely eligible, and refuse to mass-create items for the ones who aren’t — because a wall of unsourced items gets deleted and the account flagged. Eligibility is earned; the system just makes sure we never miss it once it’s been earned.
Why Wikidata, and why on a schedule
A Wikidata item is the machine-readable identity card Google’s Knowledge Graph and large language models reach for when they need to know a person is a real, distinct entity. It’s what disambiguates you from the five other people who share your name, and it’s increasingly what AI answers cite when they describe you. But a brand’s eligibility changes over time — a new book ships, a company gets acquired, an athlete earns national press. A person who wasn’t eligible last quarter may be eligible today. A one-time audit goes stale the week after it’s run. So we made it recurring.
The four tiers every name falls into
The audit sorts every brand into one of four buckets. The discipline is in being honest about which bucket someone is actually in.
Tier A — Already on Wikidata → link & enhance
Independent coverage already earned them an item. Their job isn’t creation — it’s making sure the item points to their entity home (the official website statement), that the description is accurate, and that the website points back. We close that loop. Eight fleet brands are here today, including Dennis Yu, Daniel Larimer, Julian David, and Nathaniel Stevens.
Tier B — Eligible, no item yet → create on a drip
A verifiable company plus a book or independent press means they can be described with serious sources. We create these one or two at a time, each with real references, after confirming no existing item already covers them and no famous namesake will be confused for them.
Tier C — Build notability first (the largest group)
Real, talented people whose coverage is still mostly their own channels — emerging athletes, local-service owners, creators. A Wikidata item created now would be deleted. Their play is the entity home + Person schema, which has no notability gate, while they earn the coverage that makes an item stick later.
Caution — Namesake traps → never link
Some brands share a name with a famous, unrelated entity — a diplomat, a surgeon general, a pro athlete, even an item literally labeled “email spammer.” We flag these and never link a client to a wrong or negative entity. A misleading authoritative reference is worse than none.
The rule we won’t break: don’t mass-create
The fastest way to lose Wikidata presence is to try to manufacture it in bulk. Wikidata is heavily patrolled; unsourced or promotional items get nominated for deletion within days, and a young account that creates a roster of borderline items at once can get flagged — taking the legitimate items down with it. So we drip: a small number of well-referenced items per week, each defensible on its own. Patience here is the strategy.
How the weekly system actually runs
Each week the audit does the same five things, reading from a living registry so it stays fast and never re-derives what it already knows:
- Pull the live fleet and diff it against last week — any new site shows up automatically.
- Re-verify every existing item — is it still live, still accurate, still pointing to the entity home?
- Enhance — add a missing official-website link, a precise description, or a reference. (This week that meant adding the website link to the Larimer and Julian David items, and a clean English description to Larimer.)
- Create from the queue — one or two newly eligible people, references verified first.
- Record everything — the registry and audit log are updated so the next run starts where this one left off.
Where this fits the bigger picture
Wikidata is not step one. It’s the payoff of a sequence we score on the 100-point Personal Brand Score: entity home first, independent authority second, Wikidata once you’ve earned it. The full playbook lives in our Wikidata & Knowledge Panel series.
The full Wikidata & Knowledge Panel series
→ Who is eligible for a Wikidata page? (the notability test)
→ How to use Wikidata for a Knowledge Panel (the build steps)
→ Earn a Knowledge Panel with a common name
→ Should you add a middle initial?
Want an honest read on your Wikidata eligibility?
We’ll tell you the truth — eligible now, borderline, or build-first — and then do the work: entity home, references, and the item once it’s earned.