AI Reputation Audit
The AI reputation audit is the diagnostic everything else starts from. We ask every major AI assistant the questions your buyers, investors and recruiters actually ask, record every answer in every mode, trace each claim to the page that produced it, and grade what can be removed, corrected or outranked. You get the full picture in one private report, before committing to any repair work.
Tell us the name to check and where you operate. A specialist scopes the prompt set on a short call and delivers the full report privately.
Something went wrong sending your request. Please try again, or WhatsApp us directly.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
What Is An AI Reputation Audit?
Most owners have never once asked an AI assistant about their own name, let alone asked it the way a sceptical buyer would, across five assistants, in every mode, repeatedly. So the first honest step in AI search reputation management is not a fix. It is a measurement. The audit establishes, with evidence, what each assistant actually says about you today, where every damaging sentence came from, and what it would genuinely take to change it.
That last part matters. AI reputation work has attracted its share of vendors selling guarantees they cannot keep. An audit-first engagement protects you from that: before any retainer, you see the exact answers, the exact sources, and our honest grading of what is removable, what is suppressible, and what is simply old and needs to be outranked. If your answers are clean, the report says so, and you walk away with a baseline and a monitoring option instead of a project you did not need.
One Question Is Not An Answer
Assistants vary their answers between runs, between modes and between surfaces. A single screenshot, good or bad, proves almost nothing. The audit runs a structured prompt matrix: the questions buyers really ask, crossed against every assistant, in browsing and non-browsing modes, repeated over days. Patterns are what we report, because patterns are what your buyers statistically encounter.
The Exposure Matrix
The audit's centrepiece: your real-world questions crossed against every assistant, each cell graded from repeated runs. A composite example for a fictional clinic shows how localised the damage usually is.
| ChatGPT | AI Overviews | Gemini | Perplexity | Copilot | |
|---|---|---|---|---|---|
| "Northbridge Dental" | ✓ | ✓ | ✓ | ✓ | ✓ |
| "...reviews" | ! | ✗ | ! | ! | ✓ |
| "is ... legit?" | ✗ | ! | ✓ | ✗ | ! |
| "...complaints" | ! | ! | ! | ✗ | ! |
| "...vs competitors" | ✓ | ✓ | ✓ | ✓ | ✓ |
Damage is rarely everywhere. It concentrates in specific question shapes on specific assistants, which is exactly what makes it fixable once it is mapped.
From Answers To A Graded Source Map
Capturing answers is the easy half. The audit's real work is tracing every damaging sentence to its page and grading what can be done about it.
A graded source map. Every offending URL marked removable, correctable, suppressible or outdated, with the repair plan, timeline and honest odds for each.
The graded map is the deliverable the whole engagement is built on. It is also how you hold us, or any vendor, accountable for what was promised.
Roughly one audit in five ends with "you have no AI problem worth paying to fix". You get the baseline, the evidence, and a monitoring option so you hear when that changes. An audit that always finds a crisis is a sales document, not a diagnostic.
What The Report Contains
A private document, typically delivered within two weeks of scoping, built to be acted on.
The Full Answer Record
Every prompt, every assistant, every mode, with the exact answers captured and dated, so you see precisely what a buyer, investor or recruiter meets today, and you have the before picture for any work that follows.
The Graded Source Map
Every damaging claim traced to its source pages, each graded removable, correctable, suppressible or outdated, with the specific route we would use: platform reporting, publisher outreach, de-indexing, entity correction or suppression.
The Machine-Readability Check
How your own website reads to AI crawlers: factual clarity, structured data, entity consistency across profiles, and llms.txt readiness, with the fixes ranked by effort against impact.
The Plan With Honest Odds
A prioritised repair sequence with per-source timelines and realistic probabilities, plus what we would monitor afterwards. Fixed scope, no obligation to continue past the audit.
How The Audit Runs
Six stages, typically two weeks end to end. You are involved at the start and the finish, and left alone in between.
Scope the name
A short call to pin down name variants, markets, languages and the awkward questions your sales team already hears. Individuals and companies get different prompt shapes.
Build the prompt matrix
Your real-world questions crossed against ChatGPT, Google AI Overviews and AI Mode, Gemini, Perplexity and Copilot, in browsing and memory modes where each applies.
Capture over days
Every cell is run multiple times across several days, because answers vary and single screenshots lie. Citations are logged wherever an assistant shows them.
Trace and grade
Damaging sentences are traced to their source pages and graded for removability, with the Perplexity citation trail used as the cross-assistant source map.
Write the report
Findings, the graded source map, the machine-readability check and the prioritised plan, written in plain language you can hand to a co-founder or a board.
Walk it through
A private session to go through the report, answer questions, and agree what, if anything, happens next. The report is yours either way.
Find out what every AI says about you, in one report
One scoping call, two weeks, and you know exactly where you stand across every assistant, with the evidence to back it.
When To Run An Audit
The audit earns its cost fastest in these moments.
Before a raise or exit
Investors run AI diligence now. Knowing what Deep Research says about you before they do is the difference between managing a narrative and reacting to one.
After bad press or a dispute
The matter may be closed, but the assistants may not know that. The audit shows whether the resolution is retrievable or the old story still leads.
Before spending on brand
A campaign that drives people to ask AI about you is wasted if the answer is wrong. Audit first, then promote into a clean answer.
Or simply because you never have
Most names have never been checked. One audit establishes the baseline everything else is measured against.
Start my audit →AI Reputation Audit Questions
What exactly do I get, and how long does it take?
What do you need from me to run it?
Is the audit confidential?
What if the audit finds nothing wrong?
Which assistants and modes are covered?
How is the audit priced?
Is this a one-time thing or ongoing?
How is this different from just asking ChatGPT about myself?
Know where you stand with every AI, before it costs you.
One scoping call starts it. Two weeks later you have the answers, the sources, the grades and the plan. Private report, honest odds, no obligation.