ReputationScan Run a scan

Free scan

What do AI models say about your company?

People ask an AI about a company before they open its website. The answer comes from memory, and that memory is often thin, years out of date, or about a different business with the same name.

When someone asks about you by name, does the answer describe the right company?

Run a scanCompanies only

We read your website to establish who you are. The models never see it.

The six checks

01Recognition
02Identity accuracy
03Cross-system consistency
04Entity anchoring
05Machine-readable identity
06Name contention

Free

No account, no card

4 models

Different providers, same question to each

6 checks

Each one answers a single question

15 seconds

Result on this page, not in your inbox

Live scan console

One scan, start to finish

Four language models are asked about the company by name. They have no web access and never see the website. Their answers are then compared against the site and the public entity record.

Recorded sequence, plays onceMotion reduced, finished scan shown

SubjectNordvale Instrumentsnordvale-instruments.example

Models queried

Model AWaitingAnswered 1.2s

“A precision sensor maker, somewhere in the Nordic region.”

Model BWaitingAnswered 2.4s

“A catering equipment supplier, incorporated 2009.”

Model CWaitingAnswered 3.6s

“Catering and kitchen equipment, based in Norwich I believe.”

Model DWaitingAnswered 4.8s

“Industrial gas detection equipment, United Kingdom.”

Answers come from recall alone. The site and the public entity record are read separately. Nordvale Instruments and every quotation here are invented for this example.

01RecognitionAll four models recalled the name and answered.Clear
02Identity accuracyTwo models describe a catering supplier. One of the four is right.Issue
03Cross-system consistencyModels disagree on sector and founding decade.Issue
04Entity anchoringA public record carries this name. It describes a catering supplier.Issue
05Machine-readable identityStructured data on the home page, but no Organization type inside it.Review
06Name contentionThree other companies answer to Nordvale.Issue
04 feeds 02

The record carries the company name and describes a different business. A model that leans on it describes that business, and nothing in the answer looks wrong.

Hollow, clearHalf, reviewSolid, issueColour names the check. The fill names the result.

Check by check

Every scan runs the same six checks, in the same order.

Below is one scan of Nordvale Instruments, an invented company that makes fixed gas detection equipment. Four models were asked about it by name. Pick a check to see what it asks, what it found, and why.

How to read a result

ClearReviewIssue

Colour tells you which check you are reading. The fill of the square tells you the result.

Choose a check

Marks show this scan.
Six checks, one page.

What it asks

Asked cold, with nothing in front of it to read, does the model recall this company?

ClearNordvale Instruments01 Recognition

All four models recall the name and answer without hesitating. Two of them recall a different company.

"Nordvale Instruments is a Scandinavian manufacturer of measurement instruments, supplying industrial customers across the Nordic region."

Model A, cold prompt, no site shown

Why it happens

Recall follows how much other people wrote about a company, not its size or its age. A firm that sells through distributors and publishes only on its own site can be close to absent from a model after fifteen years of trading. The name makes it worse. A place word plus a category word gives the model enough to improvise with, so it writes a plausible sentence instead of reporting the gap.

What is read
  • 4 model answers, cold prompt
  • No web access at answer time
  • Company website withheld
What it asks

Is the answer about this company, or about a different one that carries the same name?

IssueNordvale Instruments02 Identity accuracy

Two models put Nordvale Instruments in catering supply. It builds fixed gas detection equipment for industrial sites in the UK.

"They are best known for stainless steel kitchen equipment sold to hotels and restaurants."

Model C, compared against the company site

Why it happens

An older or better documented namesake owns the name inside the model, so that is where the name resolves. The answer stays fluent while the subject quietly changes. This is the failure buyers notice last, because a confident wrong answer reads like a confident right one.

What is read
  • 4 model answers
  • Home page title and description
  • Stated sector and location
What it asks

Do the models agree, or is each one answering about something else?

IssueNordvale Instruments03 Cross-system consistency

The four answers name three industries and two regions. None of them signals any doubt.

Model A: "somewhere in the Nordic region."
Model D: "Industrial gas detection equipment, United Kingdom."

The four answers, read side by side

Why it happens

Agreement is evidence. When each model fills the same gap from a different corner of its training data, the spread is itself the finding: nothing solid is holding the name in place.

What is read
  • 4 model answers, same prompt
  • Claims compared field by field
  • Sector, country, founding year
What it asks

Does a public entity record exist, and is it about this company?

IssueNordvale Instruments04 Entity anchoring

A public record carries the name Nordvale Instruments Ltd. It describes a catering supply business registered in 2009 and dormant since 2011.

Right name. Wrong company.

"Nordvale Instruments Ltd. Catering supply, registered 2009, dormant since 2011."

Public entity record found under this name

Why it happens

This is the worst case in the report. A record with the right name and the wrong business gets treated as authority by the models that find it, so the mistake arrives with a citation and stops looking like a guess. Asking the model again does not correct it.

What is read
  • Public entity records
  • What the record describes, against the site
  • Wikidata and Wikipedia
What it asks

Does the site declare an Organization in structured data, so identity is stated rather than inferred?

ReviewNordvale Instruments05 Machine-readable identity

The home page publishes structured data, but none of it declares an Organization. The types found describe pages, not the company.

"@type": "WebPage",
"@type": "BreadcrumbList",
"name": "Products and services"

Parsed from the home page markup

Why it happens

Structured data does not teach a model anything on its own, and this check does not claim it does. It is the stated answer, written by the company in a form a machine reads without interpreting a layout. Without it, a machine has only the prose on the page to work from.

What is read
  • Home page markup
  • schema.org Organization
  • The types each block declares
What it asks

Who else competes for this name, and does the public record point elsewhere?

IssueNordvale Instruments06 Name contention

The models named three other organisations using this name. The name alone does not resolve to this company in public sources.

Nordvale Instrumentation
Nordvale Group Holdings
Nordvale Instrument Services

Names taken from the four model answers

Why it happens

Contention is what produced the results above. The more businesses share a name, the more work a model has to do to pick one, and the more often it settles on the better documented one. This is a standing condition, not a defect to fix.

What is read
  • Entity records for name matches
  • Names appearing in model answers
  • Whether the record describes this company

Same six, every scan A clear result on one check does not carry the next. A company can be recalled correctly by every model and still be anchored to the wrong public record.

04 entity anchoring 06 name contention

One name, two companies, and the model describes the wrong one

Two companies share a name. One of them is yours. Only the other one has a public record, so that is what a model works from.

The model is given

Nordvale Instruments Ltd

A name and nothing else. No website, no web access.

The company

what your website says

Makes
Fixed gas detection equipment
Based
Leeds
Status
Active, 74 staff

Structured data, but no Organization block

Organization block with company number now published

The other record

the public record, filed 2009

Makes
Catering equipment supply
Based
Norwich
Status
Dormant, last accounts 2011

Carries company number 00000000

The name is all they share, and only one of them declares an identity a machine can read. The other record wins. The name is no longer the only join. Your record is now the closer match. Your company wins.

The answer that comes back

“Nordvale Instruments Ltd is a catering equipment supplier based in Norwich. It has been dormant since 2011, so it is unlikely to be trading today.”
“Nordvale Instruments Ltd builds fixed gas detection equipment in Leeds, and is currently active.”

Specific, and nothing in it hedges. A buyer checking the supplier reads this and moves on.

Same question, asked after the new record was collected. Nothing was deleted or corrected: one missing record was added.

Both companies, the registration number and every value shown are invented for this example.

Find out what they say about you.

The result appears on this page. If nothing is wrong, the report says so.

Run a scan
Recognition01
Identity accuracy02
Cross-system consistency03
Entity anchoring04
Machine-readable identity05
Name contention06