A Knowledge Panel is one of the clearest signals that Google — and by extension, AI tools trained on the same data — actually recognizes who you are.

This guide walks through the practical steps to build the entity signals that make a Knowledge Panel possible, whether for a personal brand or a company.

A Google Knowledge Panel is the box that appears on the right of the search results when Google is confident it knows who or what something is. It is not a design element and it is not something you can buy. It is Google telling you it has assigned your name an identifier in its Knowledge Graph and attached a confidence score to it.

That distinction matters more than anything else in this guide. Most advice on getting a Knowledge Panel treats it as a checklist of things to add to your website. It is not. A panel is the output of a decision Google makes about you, and your website is only one of the inputs.

Google mints a panel when two things are true at once

The first is that Google can tell exactly who you are — one name, one identifier, one consistent set of facts, distinguishable from everyone else who shares your name.

The second is that enough people are asking. Google does not spend effort disambiguating entities nobody searches for.

Your website can solve the first problem completely. It can do almost nothing about the second. Most people fail because they spend all their effort on the half they control and none on the half that decides the outcome.

Step 1: Pick one name and never vary it

Decide the exact string you want to be known by, down to the punctuation, and use it everywhere. Not “Rob” in one place and “Robert” in another. Not “Smith Consulting” and “Smith Consulting Pty Ltd”.

Corroboration works partly by string matching. Five slightly different versions of your name read as five weakly-related people. One repeated version reads as one person, confirmed five times.

Write a single biography of about forty words in the third person and freeze it. That exact text goes on your site, your social profiles, your author pages, and every bio you are ever asked to supply. Third person is not a style choice — a sentence beginning “I” cannot be bound to an identifier, so first-person copy produces no usable facts at all.

Step 2: Build an entity home

An entity home is the single page on the internet that is the definitive source of truth about you, and the page every other record points back to. For a person, it is almost always the root of a domain that matches your name.

The page needs to state, in plain third-person sentences, the things you want Google to know: what you do, where, what you have published, what you are known for, who you work with. Structure it the way an encyclopedia entry is structured, because that is the format entity extraction is tuned on.

Then mark it up with structured data — a Person node with a stable identifier, connected to nodes for anything else you own: a book, a podcast, a company, a series. The identifiers are what turn a collection of separate facts into one entity with relationships.

Step 3: Connect yourself

The sameAs property is a list of other URLs that are also you. It is the cheapest high-value thing on this list and the one most often left half-finished.

Include every profile that genuinely represents you: social accounts, author pages, professional directories, podcast platforms, company listings. A sameAs array with one entry is doing none of its job. Aim for a dozen or more.

Step 4: Corroborate — the step that actually decides it

Everything above is self-reported. You wrote it, on your own site, about yourself. Google discounts it accordingly. What moves the needle is independent sources saying the same things.

Not all sources are equal. What counts most are authority records — structured entries in databases that exist to identify things, maintained by institutions with a reason to be accurate.

If you have published a book, you have access to a set of these that most people do not:

National library deposit. In Australia, depositing a publication with the National Library is a legal requirement for publishers, including self-publishers. It is free, and an electronic deposit produces a National Library catalogue record and a Trove record within roughly forty-eight hours. That is a government-backed, machine-readable, third-party record about your work — and it is the fastest one available.
ORCID. A free, permanent identifier for anyone who authors work. Not restricted to academics. Takes about ten minutes.
ISNI and VIAF. Library-grade person identifiers that follow on from the records above.
Open Library and Google Books. Both free, both mint identifiers that other databases accept.

Without a book, the equivalents are professional registries, company registers, industry directories with editorial standards, and genuine press coverage. Guest appearances on established podcasts are among the most underrated: each one produces a page on a domain you do not own, carrying your name and your bio.

Step 5: Wikidata goes last, and here is why

Wikidata is the most direct line into the Knowledge Graph, which is exactly why people rush it and lose.

An item is accepted if it has a Wikipedia article in any language, if it refers to something described by serious publicly available references such as a database or authority record, or if it fills a structural need in the data model.

Items get deleted when they have no external references, when the description reads as promotional, or when the account creating them appears to represent the subject. An item about yourself, made by yourself, sourced only to your own website, is the textbook deletion case — and a deletion leaves a record against the name that makes the second attempt harder than the first.

Do steps one to four first. Then create the item for the work — the book, the company — cited to the authority records you have accumulated, and only then a person item whose claim is that they authored it.

Step 6: Give people a reason to search for you

This is the half nobody writes about, and it is the half that decides the timeline.

Publishing, speaking, appearing on other people’s platforms, and being mentioned alongside people who are already entities in your field all do the same job: they generate the branded search volume and the co-occurrence signals that make disambiguating you worth Google’s effort.

A perfectly marked-up person nobody looks for does not get a panel. Someone with messy markup and real demand often does.

How long it takes

For an established public figure with press coverage, a panel can appear in weeks once the signals are consistent. For someone starting from no coverage and no authority records, six to eighteen months is realistic.

The website work is cheap and fast. The corroboration is slow and is the actual constraint. Anyone quoting a fixed short timeline is either selling schema installation or has not done it before.

Common questions

Can you pay Google for a Knowledge Panel?
No. Panels are generated from the Knowledge Graph and cannot be purchased or submitted. Anyone selling a guaranteed panel is selling the work described above, or selling nothing.

Is schema markup enough on its own?
No. Structured data helps Google read what you claim about yourself. It does not verify it. Without independent sources, a claim gains no confidence.

Do you need a Wikipedia article?
No, though one makes it substantially easier. A Wikipedia article satisfies Wikidata notability outright and is a strong corroborating source on its own.

What if someone else shares your name?
This is common and it is why consistency matters so much. Distinguishing details — a birth date in an authority record, a consistent occupation, a matching set of linked profiles — are what let Google separate you from them.

Getting help with this

Ryan Gooding builds knowledge panels and entity SEO for Australian businesses and personal brands, and writes about the subject in Branding: is bullshit. If the sequence above looks like more work than you want to do yourself, get in touch.