Assistants and answers
A thing a machine holds a record of — a company, a person, a product — as opposed to a set of pages that happen to mention a name.
Also called Knowledge graph entity · Сущность
The distinction decides what happens to an ambiguous query. Ask for a two-word company name and the machine must first decide which thing is meant, and it decides from the records it holds rather than from the pages it could return. A site with no entity behind it competes as a string; a site with one competes as a subject, and only the second can be described rather than listed.
Nothing about this is mysterious to arrange, and none of it is a trick. A consistent legal name, one canonical address for it, the same details in structured data as in the trade licence, an author who exists in more than one place, and outside records that agree with all of it. The work is agreement across sources, not assertion on your own.
It is also slow in a way that should be said before anyone buys it. Records are built from repeated, corroborated mentions over months; there is no submission form for most of them, and a new domain will spend a season being a string before it becomes a subject.
Structured data claiming one legal name while the licence, the invoice and the LinkedIn page carry another does not build an entity, it prevents one — the records cannot be reconciled, so nothing is consolidated. Fix the disagreements first and in the outside sources, not only on your own site, which is the one source a machine already discounts for being self-interested.
Knowing the definition is not the same as being able to check the figure. These are the procedures that do the second thing.
- Testing a promise about AI answers
- “We will get you cited by ChatGPT and the other assistants.” · 20 minutes, 5 questions.
The definitions are the easy part. Whether the figure on your dashboard was computed this way is a different question, and usually the more expensive one.