Definition
A knowledge graph is a database of real-world things and the links between them, holding your brand as an entity rather than a string. A search index records which pages contain a word; a graph records what a thing is and what it connects to. Assistants use them to tell a brand from an ordinary word.
At a glance
- CategoriesTechnical
- Related fieldsEntity Optimization, SEO
- DifficultyIntermediate
At a glance
- A knowledge graph stores entities, their attributes, and the relationships joining them.
- Google, Microsoft and the large AI labs each maintain their own, and they disagree.
- An entity carries an identifier, so two companies with the same name stay separate.
- Nothing you publish enters a graph directly. Graphs are built by reading the open web.
Entities, attributes and the links between them
Three parts, every time. A thing, a property of that thing, a link to another thing. Patagonia is a company. It makes outdoor clothing. It sits in Ventura, which is a city, which sits in California.
Written out, that looks trivial. The value shows up when a question needs two hops. “Which AI visibility companies are based in the Pacific Northwest” cannot be answered by matching words, because no page contains that sentence. A graph answers it by walking from a category to a set of companies to their locations.
Graphs are assembled by machines reading the web, cross-checking claims across sites and resolving which mentions point at the same thing. Google exposes a slice of its own through the Knowledge Graph Search API, a quick way to see whether your company is in there and what type it was given.
Being a thing, not a string
A search engine can rank you without ever deciding what you are. An assistant cannot answer a question about you without deciding.
That decision is where brands lose. With nothing establishing your company as a distinct entity, a model infers it from context each time, and sometimes wrongly: merged with a similarly named business, given a competitor’s founding date, filed under the wrong category.
Category is the part people underrate. If a graph has you down as a marketing agency and you sell software, you drop out of every software answer and turn up in agency answers you did not want. No notification arrives. You find out by asking an assistant what your company does.
The fix is not a submission. It is consistency, which is what entity optimization actually is: the same name, category and facts repeated in enough places that the resolution step has nothing to argue with.
Make your brand legible as an entity
Publishing work, not platform work.
- Pick one name and use it everywhere. Legal name, trading name, product names. Choose the form you want to be known by and stop varying it across your site, your profiles and your press.
- Mark up the obvious. Organization schema on your about page, Product schema on product pages. Structured data does not put you in a graph, but it removes the guesswork about what you claim to be.
- Say your category in plain words on your homepage. “Parsnipp is an AI search visibility platform.” One sentence, unhedged, in the HTML.
- Get the off-site facts consistent. Wikipedia if you qualify, Crunchbase, LinkedIn, industry directories. Mismatched founding dates and headcounts are what make a resolver hesitate.
- Connect yourself to things that already exist. Your city, your industry, the standards you support, the platforms you integrate with. Links to established entities are how a new entity gets anchored.
Frequently asked questions
-
Do I need to build my own knowledge graph?
Almost certainly not, and this is where budget gets wasted. Building one makes sense if you run a large catalog and want your own retrieval over it. For getting cited in AI search, the job is appearing correctly in graphs other people maintain, and none of that involves building anything.
-
How do I get my company into Google's Knowledge Graph?
There is no application. Entities appear once enough independent sources describe the same thing the same way. Make your name, category and core facts identical across your own site and the third-party sources that already rank for your brand, then wait for the next rebuild.
-
How long does it take for an entity to settle?
Months rather than weeks, and not on a smooth curve. An entity usually appears with missing or wrong attributes first, then corrects as more sources line up. Changing something already established, a rebrand above all, takes longer than creating it, because the old facts have to be outweighed.
-
Do AI assistants use Google's knowledge graph?
Some draw on it, some keep their own, none publish the details. The signal underneath travels across all of them: consistent public facts about a distinct thing. Optimize for that rather than for one company’s graph and you cover the ones you cannot see into.