The front door, made tangible
Sample AI Visibility Audit.
This is what the £750 Visibility Audit looks like when it lands: an anonymised example built from real audit structure, with illustrative numbers. It shows where a practice stands across the AI engines, why competitors get named instead, and exactly what gets fixed first.
What the audit shows
The Visibility Audit answers one question in measured terms: when your customers ask an AI engine who to use, are you in the answer - and if not, who is, and why? It opens with a fixed set of your buyers’ real questions run across the major engines, then works down through the layers that decide what gets cited: crawler access, entity clarity, structured data and independent corroboration. It closes with a prioritised fix list. Everything below is what that looks like when it lands.
The prompt set, engine by engine
Ten buyer questions, run across the agreed engines, each answer recorded: who was named, in what order, and which source the engine leaned on. Five example rows:
| Prompt | Named? | Who the engine recommended | Where the citation came from | Priority |
|---|---|---|---|---|
| “Best cosmetic dentist in [city]” | third | Two competitors, then the practice | A local directory profile | medium |
| “Best dental implants in [city]” | absent | Five competitors | Two directories and a regional news feature | high |
| “Invisalign cost [city]” | named | The practice | Its own pricing guide | hold |
| “Emergency dentist [city] today” | absent | Three competitors | A directory and a map-style listing | medium |
| “Is teeth whitening safe?” | no local practice | Nobody local | National health publishers | ignore |
Citation share, measured
Every run is scored into one number per business: the share of answers, across the full prompt set and engine set, in which it is named. This is the baseline every later report is measured against.
Citation share · full prompt set × engine set · illustrative
Crawler access
Google and Bing verified reachable - checked through the engines’ own inspection tools, not spoofed requests, because spoofed engine-bot tests lie.
PerplexityBot blocked with a 403 at the CDN edge - a silent block that robots.txt never shows. The practice was invisible to an entire engine and nobody knew.
fig. - the access matrix from this sample. One blocked agent removes a practice from an entire engine.
Entity consistency
The practice’s name appears in three different spellings across its own footer, its Google profile and a major directory, and no Knowledge Graph entity exists. Engines cite what they can verify; a business that presents as three near-identical entities gets treated as none of them.
Structured data
Organization schema is present but carries no sameAs links to the profiles that corroborate it, and the FAQ schema on two pages has drifted out of sync with the visible copy. Valid but unconnected schema does half its job.
The fix list
Quick wins - done in days
- Allow-list the AI crawlers at the CDN edge, then re-verify every agent
- One exact name, address and phone everywhere the practice appears
- Resync FAQ schema to the visible copy
- Answer-first capsule on the implants page - the highest-value gap in the set
Strategic - the compounding work
- Earn independent editorial mentions: the “best implants” answer is being decided by third-party citations, not on-page changes
- Establish the review-platform route so ratings surface in machine-readable form
- Measure monthly against this baseline and defend the positions won
What happens next
The report and fix list are yours either way. Most practices move into the Build to close the gaps, then the Watch to hold the position.
This is an anonymised, illustrative example. AI answers vary by engine, timing, location, personalisation, query wording and source availability - which is exactly why the audit measures a fixed prompt set against a dated baseline rather than a screenshot of one lucky answer.
Last updated: July 2026