Google Knowledge Panel — the credibility artefact you cannot buy, but you can engineer

A Google Knowledge Panel is not a product Google sells. It is a confidence threshold their Knowledge Graph crosses when enough trustworthy sources agree on the same entity facts. The work is engineering the supply of those signals — not gaming them. Done well, a Person panel triggers in 90–180 days for founders, executives, and authors with real public footprint. Done badly (generic Wikipedia attempts, fake press, paid Wikidata edits), the panel never lands and the brand earns negative signals along the way.

The clients I take on for google knowledge panel work tend to fit one of these three shapes:

• Founders and executives whose buyers Google their name before a sales call — and where the absence of a panel is itself a negative signal
• Authors, podcasters, and creators with real public footprint that has not consolidated into a single recognised entity
• Mid-career professionals expanding into thought-leadership positions (board seats, speaking, advisory) where the panel is a credibility multiplier

Wikipedia is the trap, not the goal

Most "knowledge panel agencies" pitch Wikipedia article creation. For 90% of founders this is the wrong target — the notability threshold is high, the editor community pushes back hard against promotional articles, and rejected attempts taint the entity record. Wikidata is the entity layer that actually matters; the Wikidata entry is what Google's Knowledge Graph reads, and a clean Wikidata entity can trigger a panel without a Wikipedia article at all.

sameAs is not just LinkedIn

A Person schema with five sameAs entries (LinkedIn, X, YouTube, GitHub, Crunchbase) is the floor, not the ceiling. A serious panel push extends sameAs to: podcast platforms (Spotify, Apple Podcasts artist pages), Substack / Medium author profiles, Goodreads (if you've published), Behance / Dribbble (if you're creative), conference speaker pages, board / advisory listings on Crunchbase, IMDb (if media), Google Scholar (if academic), academic ORCID. Each one is a confidence vote.

Cross-platform consistency is the silent killer

Your name on LinkedIn might be "Gautam Khorana"; on Twitter "@gautamkhorana"; on Crunchbase "G. Khorana"; on a 2018 podcast appearance "Gautam K. Khorana". Each variation costs the Knowledge Graph a confidence point. The first 30 days of the engagement is name, photo, and bio reconciliation across every public surface — the panel doesn't trigger until the entity is unambiguous.

Press mentions matter, but quality not quantity

A single feature in The Guardian or BBC moves the panel needle further than 50 paid placements in vendor publications. Google's Knowledge Graph weighs editorial outlets, recognised industry publications, and well-known podcasts. Press release services and pay-to-play "as featured in" badges are noise.

Image consistency is non-obvious but high-impact

The panel's photo is pulled from the most-linked-to public image of the entity. If your LinkedIn photo, Twitter photo, Crunchbase photo, and About page photo are all different shots, the Knowledge Graph hesitates. A single canonical headshot used across every public surface is the fastest single move.

• Entity audit — what Google's Knowledge Graph already knows about you, and where the contradictions are
• Person schema rollout — full Person markup on the homepage and About page with sameAs covering 12+ platforms (not just 5), worksFor / alumniOf / knowsAbout / nationality / homeLocation
• Wikidata entity creation — clean, well-cited Q-item with statements that mirror the schema.org Person facts
• sameAs unification — single canonical name spelling, single canonical photo, single canonical short bio, propagated across every platform that exposes a sameAs surface
• Press / podcast outreach — 6–10 strategic placements per quarter on outlets the Knowledge Graph reads (industry trade publications, named podcasts, conference speaker pages)
• Author bylines on every blog post the person publishes, with structured Article + Person schema linking back to the canonical Person entity
• Monthly tracking — entity health, panel surface check (does it trigger? is it complete?), recommendation queue for the next 30 days
• Quarterly review and 90-day re-targeting based on what the Knowledge Graph has actually picked up

Project range: £4–12k for the build, depending on PIM scope, multi-language requirements, and the size of the existing catalogue. Ongoing retainer optional.

I run a WordPress agency with 5,000+ sites at Seahawk Media — that is the credibility on the agency side. The interesting half is the gemstone domain knowledge: I built and maintain the largest catalog of astrological gemstone content on the public internet at Deluxe Astrology — 125+ gemstones, 30 languages, full planetary and astrological pairing logic. I know what a Pukhraj is, what GIA cert numbers look like, what Rapnet feeds carry, and how to render a parcel-pricing table that the trade actually wants to use.

The retail-jewelry agencies (GemFind, Jewelry Website Designers, Lounge Lizard, WebFX) build Shopify storefronts for B2C jewelers and have done so for 20+ years. The generalist Indian dev shops (Magneto IT, TransPacific) ship volume work but do not speak the language. The intersection of "modern stack" and "actually understands the gemstone trade" is small enough that I am the only person on it I am aware of.

• GEO / AI SEO The broader AI-search-visibility programme — entity work overlaps, but the buyer intent is different.
• Jewelry SEO Where the same entity-graph work supports a brand-led SEO programme rather than a personal one.

You book the 30-minute call, you describe the business, the catalogue size, the timeline. I tell you whether I am the right person; you walk away with a stack pick, a price range, and a realistic delivery window. No deck, no qualification screen.