How a business becomes unambiguous to machines

What identity means to a machine

There is no registry a model consults. Identity is inferred: several sources describe something, the descriptions overlap, and the overlap becomes an entity. Where they do not overlap, the machine either guesses or creates two entities out of one business.

This is why the two required fields matter so much. In Google's local business structured data, name and address are the required properties, with opening hours, telephone, geo coordinates and price range recommended on top. The requirement is not bureaucratic. Those are the two things that let a machine distinguish you from a business with a similar name in the next town.

The three ways businesses split themselves in two

Legal form drift. Müller GmbH on the website, Mueller GmbH in one directory, Müller Bedachungen on the invoice. Three names, three candidate entities.

Address abbreviation. Hauptstraße 4, Hauptstr. 4 and Hauptstrasse 4 are three strings. A human reads one address, a matcher sees three.

Moving. The old address stays live on four listing sites for years, so half your evidence points at a building you left in 2019.

What Google requires and what it merely recommends

Property	Status	Why it matters for identity
name	Required	The primary handle a machine matches on
address	Required	Separates you from similar names elsewhere
telephone	Recommended	A second matching key, if formatted consistently
openingHours	Recommended	Answers a question directly, so it gets quoted
geo	Recommended	Resolves ambiguity when the street name repeats

Status column follows Google Search Central on local business structured data. Recommended properties improve understanding but are not required.

The repair, in the order that pays

Write down the canonical form once: exact legal name, exact street spelling, one telephone format. Then correct the sources you control, then the ones you can claim, then ask for the rest. Most businesses find between four and eight places carrying an old variant.

For scale: 45 percent of consumers used AI tools to find a local business in 2026, up from 6 percent in 2025, and ChatGPT alone carried 74.78 percent of AI referrals. Ambiguity now costs across 4 or 5 platforms at once rather than one.

The scale is usually smaller than people fear. A typical local business appears on roughly 8 to 15 sources: its own site, Google, Bing, Apple Maps, two or three trade directories, a chamber listing and a few aggregators. Of those, an owner controls perhaps 3 and can claim another 4. That leaves a handful to email, which is why the whole job fits in a day rather than a quarter.

Do not expect markup to fix it. Google states that structured data does not guarantee that features consuming it will appear, so schema clarifies what your page means but cannot overrule what six other pages say. Agreement is the mechanism, and agreement is edited by hand.

Questions and answers

Does the legal suffix have to be in the name?

It has to be consistent, which usually means including it everywhere or nowhere. Mixed use is what creates two entities out of one business.

How many sources need to agree?

There is no published threshold. Practically, two agreeing sources beat five that differ, because disagreement forces a guess and agreement does not.

Will schema markup fix inconsistent data?

No. It describes one page accurately. It cannot correct what other sites say about you, and Google states markup guarantees nothing on its own.

What if a directory refuses to update my old address?

Then the entry is working against you and removal is better than correction. An entry pointing at the wrong building is worse than no entry.

Sources

Google Search Central, Local business structured data Required and recommended properties, and the guarantee disclaimer
BrightLocal, Local Consumer Review Survey 2026 1,002 consumers; the share using AI for local search
SE Ranking, AI traffic research study, 18 June 2026 101,574 websites; platform shares