For years, SEO advice was about keywords: figure out what people type, put those words on your pages. That mental model is now badly out of date. Google and AI assistants do not primarily track words anymore. They track things: specific people, specific businesses, specific places, and the relationships between them. The industry word for these things is entities.
Here is the difference in practice. Old model: your page contains “web designer Montréal” so it might rank for that phrase. Current model: the machine knows there is a business called Studio Squalli, run by a person called Ahmed Squalli, located in Montréal, offering web design, related to certain client sites. When someone asks for a web designer in Montréal, the machine looks through the entities it knows and picks the ones that fit. If your business exists in its map of the world, you can be picked. If you are just words on pages, you mostly cannot.
How machines build an entity
A machine becomes confident an entity exists the same way a cautious person does: by hearing about it from several independent places that all agree. Your website says who you are. Your Google Business Profile says the same name, same city, same services. Your LinkedIn matches. A directory matches. A client’s review names you. Every consistent mention is a stitch, and enough stitches make the entity solid.
The reverse also holds, and this is where most small businesses hurt themselves. The website says “Smith Consulting Group”, Google shows “Smith Consulting”, LinkedIn says “J. Smith Advisory”, the addresses differ, and one profile still lists services abandoned years ago. To a machine, that is not one solid entity; it is noise. And assistants deciding who to cite skip noise, because a machine will not vouch for something it cannot pin down.
The work, concretely
Decide the canonical facts once. Exact business name, one-sentence description, services, city, service area. Write them down. This sounds trivial; almost nobody has done it.
Say them plainly on your own site. Your homepage and about page should state who you are, what you do, for whom and where, in ordinary sentences. Your site is the root source every other mention gets checked against.
Make every external profile match. Google Business Profile, LinkedIn, directories, social accounts: same name, same description, same details, character for character where possible. This costs an afternoon and it is one of the strongest returns on an afternoon in this whole field.
Add the machine-readable layer. There is a format for declaring, directly in your site’s code, “this is a business, this is its name, its founder, its location, its services, its prices”. It is the closest thing to filling in the machine’s form yourself instead of hoping it guesses right. I covered it in the structured data guide.
Get mentioned where you do not control the text. Reviews that name your business, a client’s site linking to yours, a local article, a podcast note. Independent mentions are worth more than anything you say about yourself, precisely because you did not write them.
Why this compounds
Keyword rankings are rented; they shift with every algorithm update and every new competitor. An entity is owned. Once machines reliably know who you are, every new page you publish, every review you earn and every mention you pick up attaches to it and makes it stronger, and both Google and AI assistants draw on the same map. Local businesses especially feel this, since local visibility is entity matching almost by definition.
Making a business legible as an entity is baked into everything I build at Studio Squalli. The free audit includes a look at how machines currently understand your business, which is often a surprise to the owner.