Wire at a dial console, Murmur in headphones, Fold holding up a circled answer card and Pip on the ladder of a great brass listening dish on a rooftop at night, beside open dark sky

Visibility You can start this yourself today.

Hear what the AI engines say about you.

See what ChatGPT, Claude and the other AI engines say about your brand, and what they get wrong. The first audit is free.

The example

What the AI engines say about your brand, what they get wrong, and how to fix it.

From Loverboy's published reading: what one engine said, what the record says, and the line Jinn drafts to set it straight.

Before

Example brand: LoverboyEngine answer

What Grok said about Loverboy

Kept as the engine gave it.

“the hard-kombucha/seltzer brand founded by Vanderpump Rules star Tom Schwartz... It sells canned alcoholic beverages”

Grok, in Loverboy's AI Brand Index reading, published July 14, 2026.

Check

Example brand: LoverboyChecked against the record

  • WrongVanderpump Rules star Tom SchwartzThe record: Kyle Cooke of Bravo's Summer House founded Loverboy. Tom Schwartz is unaffiliated with the brand. Sources in the reading: forbes.com, en.wikipedia.org.
  • RightIt sells canned alcoholic beveragesTrue. Loverboy sells sparkling hard teas and canned cocktails.

Jinn checks a claim against your own record. It does not decide whether a claim is legally allowed; your reviewer still does.

Each line of the answer is weighed against the record: one wrong, one right.

Fix

Example brand: LoverboyAbout page line, draft

Who's behind Loverboy

A short line for the About page, written from the record.

Loverboy was founded by Kyle Cooke of Bravo's Summer House.

We make sparkling hard teas and premium canned cocktails, crafted for good times.

The fix goes where engines read: a plain line on the brand's own About page.

Approval

Example brand: LoverboyWaiting on the brand

  • About page lineMatches the record

Waiting for a yesBrand lead

The fix waits for someone on the brand's team. Loverboy is an example brand, not a customer.

See the full reading

At a Glance

What Visibility covers.

Measure

  • The audit
  • Engine coverage
  • Brand IQ math
  • Knowledge Check

Learn more

Diagnose

  • Gap cards
  • Authority gap
  • Prompt intelligence
  • AI readiness

Learn more

Fix

  • Corrections
  • Playbook
  • Drafted assets
  • Content calendar

Learn more

Watch

  • Monitoring
  • Alerts and inbox
  • Trends and tracking
  • Weekly digest

Learn more

Prove

  • Proof
  • Board report
  • Brand Strategy in a Box
  • The evidence register

Learn more

Try it free

What the AI engines say about your brand.

"Buyers ask the AI engines about you, so make sure the answer is right."

Biscuit, the Goldendoodle (dog) Uwee, inviting you to start your own brand record
Biscuit, Chief Customer Companion

Watch It Work

One AI answer, checked against the facts.

The question

Jinn puts your real buyer questions to the AI engines and keeps each answer as it came back.

ChatGPT, Claude, Perplexity, Gemini, Grok and Google AI Overviews all get the question, and each answer is saved word for word, with its date.

The wrong line

Jinn checks each answer against your brand record and flags the line that is wrong.

In this example, one line is marked wrong and one right, each with the reason from the brand record beside it.

The fix

Jinn drafts a correction for each wrong line, and it waits for your yes.

Each correction is written in your brand's own voice, for your About page, your FAQ or your playbook.

The second ask

Stale, the old CRT monitor, on the observatory desk beside Wire's rotary telephone, a question card on its glowing screen and a desk calendar with one page turned. Stale, Time Keeper & Dates Editor, asks again.

After you approve a fix, Jinn puts the same question back to the engines.

You compare the new answer with the first one, and the fix you confirmed stays in your brand record.

How It Works

Ask, check, score, fix, confirm.

Start it yourself today with the free audit, and you decide what changes.

  1. Ask. Your real buyer questions go to each engine.
  2. Check. Each answer is kept as it came, with the date, and checked against your brand facts.
  3. Score. You get a Brand IQ readout and a ranked fix list.
  4. Fix. The fixes are drafted from your brand record and the corrections you confirmed. Your brand record is how your brand looks, how it sounds and what it is allowed to claim, read from your own site and confirmed by you.
  5. Confirm. You confirm or reject each correction yourself. Once you confirm a fix, it goes into the record.

See how the audit works

"Here's what the engines are saying. Here's what's wrong."

Wire turning the crank of a big brass clockwork machine, cards glowing in its slots under a listening dish
Wire, GEO & AEO Engine Reporter

AIs are already talking about you and getting it wrong.

We put real buyer questions about famous brands to the AI engines, then checked each answer against the facts, claim by claim. The card shows how often each engine got a fact wrong or out of date.

The research

The July 2026 study

Two buyer questions about each of ten famous brands, put to six AI engines: 120 answers, captured July 9, 2026, then checked claim by claim.

  1. Across the six engines in this study, 54 of 126 claims were false or out of date.

Wire, the Wind-up tin robot Uwee, carrying in the study total
Each bar is one engine’s claims checked; the filled part is the claims found false or out of date. Claims-checked counts differ per engine, because engines volunteer different amounts of checkable material. The error rates are not sample-size matched.
Each of the six engines in this study, the July 2026 study: claims checked, and found false or out of date
Engine, as the study names itClaims checkedFalse or out of dateShare
Claude (no web)191789%
Grok201575%
Perplexity201470%
ChatGPT (search)32413%
Gemini 2.5 Pro17212%
Claude (web)18211%
All six engines in this study12654

Source: the July 2026 study. Every flagged claim, word for word, is in the annex.

How the study was run

Under the Hood

How the score is worked out

  1. Scale

    Prominence is scored on a published ladder from 0 to 5, by text rules: where your brand appears in the answer, and whether the answer recommends it. How results are measured

  2. Arithmetic

    Brand IQ is arithmetic, not opinion. Each engine's score is weighted and summed to one number from 0 to 100. The same answers always give the same number. How results are measured

  3. Models

    Monitoring always uses the smallest class, even on the top tier. A higher tier buys more samples and more engines, not a bigger model. How models are chosen

  4. Fix list

    Fix-list rewrites come from a fixed catalog of block types and pass two checks in code: a banned-word check and a voice rule of thumb. No model grades either. How the check works

Let's dive deeper.

Explore why GEO and AEO are important for growing your brand, and how to set your brand up for success.

Visibility plans

A project includes

  • Your buyer questions, put to the AI engines.
  • A Brand IQ readout and a ranked fix list.
  • Corrections drafted for your playbook, FAQ and About copy.
  • A ledger of what you corrected.

You decide what changes. Each fix waits for your yes.

Start with a free audit, then talk to us. The audit asks for your email.

Visibility changes nothing on its own. You get the readout and the fix list, and the corrections go into your playbook only when you say so.

Bristle painting a jar, Ocho sorting folders, Wire on the phone and Proof checking a page, each in a lit arched window at night.

Find out what AI says about your brand. Start with a conversation.

  1. Tell us about your brand.
  2. A person on our team reads it.
  3. We read every application and reply within two business days.

Already use Claude or ChatGPT? Connect your AI, free