AI visibility audit · Findings & priorities
The AI visibility audit gives you a detailed view of how your brand appears in AI answers. I test 50 questions about your market across ChatGPT, Claude, Gemini and Perplexity, and check Google AI Overviews for 20 tracked keywords.
You'll see where your brand appears, how it compares with three competitors and where the answers contain inaccurate information. I turn those findings into a prioritised action plan, delivered with a report and a 60-minute walkthrough within two weeks of kickoff.
| Included in the standard audit | Scope |
|---|---|
| Your business | One brand and three competitors |
| AI assistant checks | 50 prompts across ChatGPT, Claude, Gemini and Perplexity |
| Google AI Overviews | 20 tracked keywords |
| Delivery | Within two weeks of kickoff |
| What you receive | An HTML report, a prioritised action plan and a 60-minute walkthrough |
| Fixed fee | £4,000 |
| Re-measurement credit | £1,000 against your first quarter of monthly re-measurement |
The free check shows what three AI models say in response to three questions. It can give you something useful to investigate, but it can't tell you how your brand appears across a whole market.
The paid audit explores a wider set of customer questions and compares the findings with named competitors. It adds checks for inaccurate information, a review of relevant technical issues and recommendations your team can act on.
The report brings the findings together with the answers and evidence behind them. You can use it to discuss priorities with your team, brief the people doing the work and return to the original results when you measure again.
An AI visibility score out of 100 summarises visibility, sentiment, accuracy and competitive standing within the audit. A breakdown by engine shows where the results differ. The score describes the tested sample, rather than every answer a customer might see.
Prompts are grouped by topic so you can see which areas need attention. The report identifies questions where your brand is absent or described less usefully than the competitors included in the audit.
For 20 tracked keywords, I record whether an AI Overview appears, whether it cites your domain and which competitor domains it cites. These checks sit alongside the four-engine results in the report.
I check the tested answers against a verified profile of your business. The report records invented facts and cases where an assistant describes another organisation under your name, with the original answer available to inspect.
You can read each prompt and the recorded answer from each engine, alongside the assessment. This gives your team a way to examine the reasoning behind the findings.
The plan sets out the changes I recommend and why I think they matter. Depending on the findings, that can include clearer business information, content improvements, structured data or other technical work. The audit also reviews schema and any llms.txt implementation, without assuming those technical details explain the results.
These anonymised examples come from two 2026 audits. They used different prompt counts, shown below, rather than the current standard scope of 50 prompts. They illustrate what an audit can find; they aren't forecasts for another business.
Across 55 prompts and four engines, one engine didn't mention the brand in any answer. Another mentioned it in six. The audit recorded 30 risk flags, including invented depot and fleet details, an incorrect founder name and confusion with a different company overseas.
The brand's domain received no Google AI Overview citations across the 20 tracked keywords. The action plan set out eight priorities, starting with clearer business information and corrections to inaccurate public claims.
This audit used 30 prompts across the same four engines and recorded 16 risk flags. Two engines described a jet-charter business when asked about the brand. Other answers included incorrect locations, invented pricing and fabricated testimonials.
Perplexity mentioned the brand in 24 of the 30 answers. The lowest mention count among the engines was 17. The findings gave the business a documented starting point for improvements and later measurement.
If you'd like to see how the findings fit together before you book, ask to see an anonymised report.
You'll have a 60-minute kickoff to explain the business and a 60-minute walkthrough to discuss the findings. I carry out the research and analysis between those sessions, with any follow-up questions needed to verify the information.
We confirm your customer groups, markets and three competitors. I use this to design questions that fit the decisions your customers make.
I run the 50 prompts across the four engines and check Google AI Overviews for 20 keywords. Discovery questions don't name your brand. Direct questions deliberately do, so we can also check how an assistant describes the business when asked about it.
I compare mentions and cited sources, check factual claims against your brand profile and review relevant technical issues. The findings become a prioritised action plan.
We walk through the report, answer your team's questions and discuss what the recommendations involve. You receive the report and plan to use with whoever will carry out the work.
You can take the plan to your own team or existing agency, ask me to help implement it or decide that other work needs to come first. The audit stands on its own.
If you'd like help putting the findings into practice, ongoing support starts at £1,500 a month. We'll agree the priorities and monthly scope around the work you need. That can include structured data, clearer brand information and content improvements, with working sessions for the tasks your team owns.
I repeat your original 50 prompts across the same four engines each month and compare them with your baseline and previous results. The report covers changes in visibility, competitor presence and newly identified inaccuracies, with relevant changes to model versions or test conditions recorded alongside the results.
Re-measurement is billed quarterly at £2,250. The £1,000 credit included with your audit applies to the first quarter, bringing that quarter to £1,250.
Additional competitors, more prompts or work across several regions can be quoted separately. We'll agree the scope and price in writing before the work starts. You can also see the wider pricing page.
Your business already invests in search, but you don't have a clear view of how it appears in AI answers. The audit gives you specific findings to assess alongside your existing marketing results.
You need to explain where AI visibility fits in the marketing plan. A documented starting point and a prioritised set of recommendations can help you decide what deserves time and budget.
A common name, a merger or overlapping product names may be making your business difficult to identify. The audit checks whether that confusion appears in the tested answers and records what needs attention.
The free check is a small sample of three prompts across three models. The paid audit covers 50 prompts across four engines, Google AI Overviews for 20 keywords and comparison with three competitors. It also includes checks for inaccurate information, technical review, a prioritised plan and a walkthrough.
ChatGPT, Claude, Gemini and Perplexity, plus the Google AI Overview checks. The report records the models and test conditions used. Results describe those tests, and may differ from answers seen in a consumer app or under different conditions.
The standard audit takes two weeks from kickoff to walkthrough. The first week covers prompt design and measurement. The second covers analysis, the report and our discussion of the findings.
You'll receive a designed HTML report with the overview, engine results, topic gaps, Google AI Overview findings, inaccurate information and prompt-level evidence. The prioritised plan and 60-minute walkthrough help you turn it into a brief for the work.
No. You can use it with your internal team, an existing agency or another specialist. If you'd like me to carry out the improvements, we agree that work separately.
The standard audit has a fixed fee of £4,000. A smaller prompt set doesn't reduce the analysis and reporting work enough to offer a lower-priced version. If that isn't the right investment yet, the free check is available as a starting point.
Use “Run my audit · £4,000” to pay by card. I'll contact you within one working day to arrange the kickoff. If you need an invoice or want to discuss the scope first, book an intro call and I'll confirm the details in writing.
Copilot isn't included in the standard audit. ChatGPT results are reported as ChatGPT results, rather than treated as a measure of Copilot. If Copilot is important to your brief, we can discuss whether a separate assessment is feasible before agreeing the scope.
You'll work with me, Kristian, the consultant behind Dog on the Table. I'm based in Brighton and run the audit, write the report and lead the walkthrough. Delivery is remote, so your team can join from wherever they work.
If you're ready to investigate how your brand appears in AI answers, the audit will give you the findings and a plan to work from. If you'd like to talk through the scope first, I'm happy to help.