Spoke
Entity consistency, and how to stop entity drift
The highest-value work in getting quoted by AI is also the dullest. Describe your company the same way everywhere, give it one stable identifier, and stop handing machines reasons to hedge about who you are.
That search systems identify things rather than match strings is well established and visible in how knowledge panels behave. That consistency specifically improves AI citation is a reasoned inference, not something any platform has published.
Search stopped being about matching text a long time ago.
"Maestro," "Maestro Strategic Partners" and a LinkedIn company page are three different strings pointing at one company. A machine that can't confidently collapse them into one thing has to hedge.
Hedging is what you're trying to eliminate. Something confident about who you are can name you. Something unsure whether the firm on your website is the firm in that directory listing will reach for whatever it's more certain about instead.
What is entity drift?
It's what happens when the same company gets described inconsistently across the web, so machines treat one company as several weaker ones. A different name in a directory, an old address on a profile, a different identifier in each site's markup — each one gives a machine a reason to doubt these are the same thing.
The damage isn't a penalty. It's dilution. Authority that should pile up against one company gets split across several partial ones, and none of them is strong enough to get named.
The audit, in the order worth doing it
Start with your own sites. This is where the worst drift hides, because everybody assumes their own properties agree with each other. They frequently don't.
Here's a real one, from the group of sites this firm runs. Three of them describe the same company, and each gives it a different identifier:
| Property | Identifier used for the same company |
|---|---|
| A consultant's site | https://consultant-site.example/#firm |
| A speaker site | https://speaker-site.example/#firm |
| This site | https://evolvelaw.ai/#maestro |
One company, named the same way, described the same way, at the same address — wearing three different identifiers. Nothing is wrong on any single page. Each is valid on its own. Together they describe a blurrier world than they need to.
Then the facts themselves. Name, address, phone, description, logo, founding date, founders, and the list of profiles you claim. Pick the official version of each, write it in one file, and correct everything against that file rather than against memory.
The address is the usual offender. "Suite 240" and "Ste. 240" are the same place to you and two different strings to a machine.
Then the profiles. Every third-party listing that says something about you is either backing up your facts or contradicting them. There's no neutral. An old office address on a directory nobody has touched since 2019 is actively working against you.
Why one identifier matters more than it looks
A stable identifier is how you say "this thing here is the same thing as that thing over there."
Declare the same identifier for your company on every property you control, and you've stated the connection outright instead of hoping somebody infers it.
The alternative — a different local identifier on each site — asks every machine to work out the relationship by matching names and addresses. Sometimes it will. The whole point of this work is to stop depending on sometimes.
The part that isn't clerical
Corroboration can't come from you.
A fact that only appears on your own site is a claim. The same fact across independent sources is evidence. That's the real reason third-party mentions matter, and it's why off-site work and entity work are the same project seen from two ends.
More in GEO & entities
How AI engines retrieve
Retrieval is what lets a model quote something you published this morning. Understanding it tells you which levers exist — and which widely repeated tactics are aimed at nothing at all.
What Google documents
Google published a guide to showing up in its AI features, and a big chunk of it is a list of popular tactics it says you don't need. That makes it the most useful source in this field, and the most uncomfortable one.
Training vs retrieval
One of these you can change this afternoon. The other changes when somebody else trains a model. Knowing which is which tells you where effort is worth spending and where it is wasted.
Platform matrix
One row per engine, each with its own last-verified date and a link to the platform's own documentation. Where a row could not be verified, it says so instead of being filled in from memory.
Comparative content
Comparison questions are among the most valuable a buyer asks and among the most common an AI engine is given. The format decides whether you can be quoted — a table in HTML can be, the same table as an image cannot.