AI Visibility Audit

Audit your AI visibility, then decide what to change first.

TendorAI’s AI Visibility Audit investigates how AI assistants answer your buyers’ questions — where your business is missing, which businesses and sources appear instead, and what to change first. If you want the agreed changes made and measured again, AI Visibility Growth is the next step.

£1,995, one-off. TendorAI is not currently VAT-registered, so no VAT is added.

Talk to TendorAI

What an AI visibility audit looks at

We ask your buyers’ questions of the assistants, record what comes back, and work from that record — measurement first, then investigation, then diagnosis, then priorities, in that order.

The measurement runs across ChatGPT and Perplexity, with five runs of every question on each assistant. Five runs rather than one because the same question does not always produce the same answer; a single run tells you what happened once, and a pattern needs repeats. The question set is agreed with you before anything is recorded, and it is then held fixed, so a later measurement can be compared with this one.

The order is the point. A priority that is not grounded in a measurement is an opinion, and there is no shortage of opinion about AI visibility. So the Audit starts by measuring where you actually stand, then works through the evidence that measurement produced, and only then says what it thinks is worth doing about it.

Every finding carries the evidence behind it. You should be able to disagree with a conclusion and still be able to check the record it came from.

Who appears instead, and which sources AI cites

When an assistant answers a question your buyer would ask, it usually names several businesses and points at several sources. Both are recorded, and they are two different things.

We record which other businesses are named in the recorded ChatGPT answers, and which sources both assistants cite. We do not run a separate measurement of each competitor, and a business appearing alongside yours is not established as a competitor.

That distinction matters enough to spell out. A business named in the same answer as yours might be a direct rival, a national brand, a directory, a supplier, or a company in an entirely different trade that the assistant has grouped with you. Which of those it is cannot be read off a recorded answer, so we report it as a business that appeared and leave the judgement to you, with the answer in front of you.

Businesses named in the same answers — read from the recorded ChatGPT answers. ChatGPT records which part of its answer each citation belongs to; Perplexity does not, so this part is ChatGPT only, and we say so rather than presenting a weaker reading as the same thing.
Sources cited — the pages both assistants pointed at, across every recorded answer, counted and grouped by domain.
Where your own domain sits in that picture — cited, retrieved but not referenced, or absent.
How your business is described when it is named, in the assistant’s own words.

A source cited in an AI response is reported as cited. That is a record of what the answer pointed at, not proof of what shaped it, and nothing on this page should be read as establishing that any page caused any answer.

Why a business chooses the Audit

Usually one of three situations. You have measured — through a Benchmark, a free report, or your own testing — and the result raised more questions than it answered. Or you have not measured at all, and you want the whole picture in one piece of work rather than in stages. Or you are deciding whether to commit to ongoing work, and you want the evidence before the commitment rather than after it.

The Audit suits a business that wants a proper evidence base before spending money on changes — including the evidence to decide that no changes are needed. That is a legitimate result, and it is one you are entitled to reach on our evidence.

How diagnosis differs from measurement

Measuring tells you what is happening. A measurement can show that your business appears in two answers out of forty, that a directory you have no presence on is cited in most of them, and that when you are named the description is three years out of date. All of that is fact, and all of it is useful.

What a measurement cannot do is tell you why, or what to do first. Those require reading across the evidence: comparing what assistants say about you with what is publicly available about you, working out which of the gaps the evidence actually supports, and judging which of them is worth the effort of closing. That judgement is the Audit.

Put simply: the Benchmark records the answers. The Audit explains them and tells you where to start.

What you receive

An agreed question set, measured across ChatGPT and Perplexity at five runs per question, and a written report built from that record. Depending on what your business turns out to need, it covers:

Your current AI visibility position, measured rather than estimated — how often your business was named, on which questions, on which assistant
How AI assistants represent your business, in their own words
Which other businesses were named in the same recorded ChatGPT answers
The sources cited in those AI responses, grouped by domain
Where your own site sits in that picture — cited, retrieved but not referenced, or absent
The public information available about you, and how consistent it is across sources
Your structured data, and what it does and does not tell an assistant
Your website and content signals
The visibility gaps the evidence supports
Findings, prioritised, each with the recorded answers behind it

The recorded answers come with it. Every finding can be traced back to the question that produced it and the answer it rests on, so you can disagree with a conclusion and still check the record it came from. How the measurement is run is set out in our methodology, and the rules we hold ourselves to are in Evidence & Standards.

What you leave with: a documented assessment of where your business stands, what the evidence shows, where the important gaps are, and what should be tackled first.

Evidence and interpretation are kept apart throughout. What was observed is stated first; what we think it means is labelled as interpretation, not presented as another fact.

Evidence, findings, actions

Three words that get used loosely, so it is worth separating them, because the Audit delivers the first two and stops before the third.

Evidence: The recorded answers themselves, and what is publicly available about your business. Facts, dated, checkable, and reported whether or not they are convenient.
Findings: What the evidence supports — where you are absent, where you are described inaccurately, where the sources being cited are ones you have no presence on. Each one carries the evidence it rests on, and each is limited to what that evidence actually shows.
Actions: What follows from the findings, in priority order. The Audit sets out what should be addressed first and why. It does not carry the work out — that is Growth, or your own team.

What a finding looks like

A worked example of the shape a finding takes, so you can see how the evidence holds it up. It is not a client result and carries no figures: it shows the chain, not an outcome.

The question
One of the agreed buyer questions, asked five times on each assistant.
The recorded answers
What came back each time, stored word for word, with whatever sources the assistant attached.
What was observed
Your business was not named in any of them. Four other businesses were named in the ChatGPT answers, and the sources cited were mostly pages published by providers in your market.
What the evidence supports
A finding that, for this question, the assistants are drawing on pages your business does not appear on. That is what the record shows.
What it does not support
Any claim about why. The record shows what was cited, not what caused the answer, and the finding says so.
What to change first
A prioritised action, with the reasoning stated, which you are free to act on with your own team.

From audit to improvement

The Audit ends with findings and priorities. It does not change anything — nothing on your website or anywhere else is touched as part of it. What happens next is your decision, and there are three reasonable answers.

Act on the findings with your own team or developer. The findings are yours, the reasoning is written down, and nothing about them depends on us.
Do nothing, on purpose. If the evidence says the gaps are not worth the effort, that is a legitimate result and one you are entitled to reach on our evidence.
Have us carry out the agreed changes and measure the same questions again. That is AI Visibility Growth.

Growth is recommended after an Audit but not required, and the Audit is complete work in its own right whether or not you go on to it. What Growth adds is the doing and the second measurement: the agreed changes made, each one recorded, then the same frozen questions asked again and compared with the baseline this Audit established.

How AI Visibility Growth works

Limitations worth knowing before you buy

The Audit is a diagnosis, not implementation. Nothing is changed on your website or anywhere else as part of it.
It is one-off work, not a programme. There is no re-measurement included, so it describes the period it was taken in.
It does not analyse your competitors. We record which other businesses are named in the recorded ChatGPT answers and which sources both assistants cite, but we run no separate measurement of any other business, and appearing alongside you does not establish that a business is a competitor.
Businesses named in answers are read from ChatGPT only. Perplexity does not record which part of an answer each citation belongs to, so that reading is not available there and is not estimated.
We do not publish a turnaround time, because we have not established one we could keep. The scope and the question set are agreed with you before the work starts.
A source cited in an AI response is reported as cited. That is an observation, not evidence that the source caused the answer.
It does not guarantee that visibility will improve, and no honest supplier can. Our own research found that businesses we changed nothing about still moved between measurements.
It covers the agreed question set and the public information we can find, not the whole internet.

None of that makes the Audit less useful. It makes it honest about what an evidence base is: an accurate description of a position at a point in time, and a reasoned view of what to do about it.

The Benchmark credit, and where the Audit sits

The Benchmark is not a prerequisite. You can start with the Audit, and plenty of businesses should — if you already expect there is a problem, paying to confirm it before diagnosing it is a step you may not need.

If you did start with a Benchmark, the fee carries over: buy the Audit within three months of your Benchmark being delivered and the £595 you paid is credited against the £1,995 Audit. The three months run from delivery. The credit applies to the Audit only — there is no Benchmark or Audit credit against Growth.

After the Audit, nothing happens automatically. The findings are yours and you are free to act on them with your own team or developer. If you would rather the work was done and measured with you, AI Visibility Growth is the next step — recommended, but not required.

See full pricing

Audit your AI visibility

Tell us what your business does and what you already know about your position. We will tell you honestly whether an Audit is the right piece of work, or whether something smaller would answer your question.

Discuss an AI Visibility Audit Compare all products