After the sprint, or on its own

What AI assistants say about you — sampled, sourced and checked.

We run a fixed set of the questions your buyers ask across agreed AI assistants, record the answers and the sources they cite, check them against your approved facts, test whether your site is technically readable, and prioritise corrections.

Request the audit See the flagship sprint

The problem

Not a score. A record of what was said, and where it may have come from.

There is a growing market in AI visibility dashboards that report a share-of-voice percentage and a trend line. Useful for tracking; not enough for deciding, because a number tells you something is wrong without telling you what to change.

This audit records the answers: what each system says about your company for the questions that produce shortlists, and which companies it names instead. We keep three things apart — citations the assistant showed, sources we infer may have influenced an answer, and provenance nobody outside the provider can see.

It also checks what is in your control. Some crawlers may miss essential information that appears only after JavaScript runs, so we test the HTML they receive, alongside access at every layer, indexing and the consistency of your company details.

  • Absent where competitors appear
  • Described inaccurately or out of date
  • Measuring for the first time
  • Re-measuring after a sprint

Deliverables

What you receive.

Answer record
What each assistant said, verbatim, by question, platform, date and run.
Cited sources, kept separate
Sources the assistant showed, listed apart from sources we infer may have influenced it.
Accuracy check
Statements compared with your approved product facts, with each error listed rather than summarised.
Competitive read
Which companies are named instead of or alongside you, and how they are described.
Readability findings
Whether essential content is in the served HTML, whether crawler access is permitted at every layer, and whether entity details are consistent.
Prioritised corrections
What you can fix directly, what depends on third parties, and when to re-measure.

The sprint includes a smaller version of this sample as its baseline. The full audit is for a deeper look, a regular re-measurement, or a company that wants the diagnosis before deciding anything else. About the sprint

Why it matters now

Absence is silent.

A company left out of an AI answer sees nothing in its analytics: no impression, no bounce, no lost-deal reason. Sampling the answers deliberately is the only way to see it.

Treat the result as a sample, not a census. It describes the questions asked, on the dates and platforms used. A change between two waves shows that something moved; it does not by itself show why.

Questions

On ai search visibility audit.

No, though the technical foundation overlaps. Classical SEO competes for a position in a list of links. This looks at how a system describes and recommends you in a written answer, where specificity and independent sources appear to matter more than ranking factors. The foundation is shared; the question is different.

No, and treat any firm that does with suspicion. These systems vary between sessions and change without notice, and nobody sells access to how they rank. What can be done is to improve the inputs — clarity, evidence, independent sources, readability — and measure whether the answers change.

Typically around twelve questions — brand, category, problem and comparison — run three times each on each agreed platform, at the same settings each wave. That is enough to see patterns and variance; it is not a survey of every buyer.

Unevenly. Systems that retrieve live pages can reflect a change within days. Answers drawing on what a model learned in training move much more slowly, and some will not change on any schedule you control.

No. The audit runs on publicly available information. Access is only needed if the work continues into changing pages.

Related answers

Questions buyers ask before they ask a firm.

How do AI systems decide which companies to recommend? What is documented, what is inference, and what nobody outside the providers knows. What is an AI visibility audit? What one should contain, what it cannot tell you, and how to judge a vendor selling one. Which AI crawlers should you allow? Every major AI crawler, what it is for, and a recommended policy — sourced from the operators’ own documentation.

See what the assistants say.

Twenty minutes to agree the questions that matter to your buyers. If an audit is the right next step, we will scope it.

Request the audit All engagements

If we can help, you leave with one specific observation about your positioning. If we cannot, we say so.

Other engagements

Where this fits.

Positioning & Evidence SprintFour weeks on one product, one use case and one buying group: a defensible position, with the evidence behind it put to work. Market & Opportunity MappingBefore the sprint: decide which market, use case and buyer to pursue first. Positioning & Category StrategyFor a broader repositioning: several products or audiences, or a decision about which category to compete in.