Methodology
One score. Every claim traceable.
A check is a recorded measurement, not an opinion. Here is exactly what gets asked, what gets stored, and where the number comes from.
Step 01
We turn your service and location into commercial questions
Not keywords — the sentences a customer would actually type into a chat. Three of them on a free check, drawn from your primary service and your main trading location, and visible before anything runs.
You see them first
The exact questions appear on the form before you submit, and you can change any of them. A number derived from questions you never saw is not a measurement you can act on.
Locations get pinned to the UK
A bare city name is ambiguous — Birmingham and Manchester exist in the United States too. We qualify the location so the engine answers about your Birmingham.
Three questions, not thirty
Each question costs real vendor spend. Three is enough to see a pattern; the rest is for monitoring, where the same questions repeat over time.
Free checks moved from five questions to three in August 2026. Scores either side of that change are not comparable.
Step 02
We ask four engines and keep what they say
ChatGPT, Claude, Gemini and Perplexity. Each answer is stored in full, together with every source it cited.
Mentioned
Were you named anywhere in the answer at all?
Cited
Was your website listed as a source the engine drew on?
The sources it trusted instead
Every page the engine cited to build its answer, ranked by how often it appears across your questions. That list is the actionable output — those pages are already trusted for your service in your area.
We do not score “was this business recommended” as its own number, and we do not discover your competitors for you. On a monitored subject we track share of voice against competitors you name; a free check reports mentions, citations and the source gap.
Gemini cites through Google grounding redirects rather than naming publishers directly — all 1,770 of 1,770 source entries in our corpus, measured 31 August 2026. We resolve them to the real publisher, because a tool that reads the URL naively sees zero Gemini citations and reports the rest as if it were the whole picture.
Step 03
We rank the sources that already outrank you
The citation list is the actionable part. Those pages are already trusted by the engines for your service in your area, so a mention on one of them is worth more than most work on your own site.
Ranked by how often they appear
A source cited across many answers carries more weight than one that appeared once.
Filtered to sources that serve the UK
A cited firm in Birmingham, Alabama is not an opportunity for a business in Birmingham, UK. We withhold any source we have not established serves the UK, and we tell you how many were withheld rather than dropping them silently.
What the score is, and what it is not
It summarises a sample: these questions, these engines, at this time.
Not a market share
No engine publishes how many people asked about your trade, so nobody can tell you what share of real conversations you appeared in. Anyone quoting you one has modelled it.
Not a forecast
These systems change their answers without notice. A score is a reading, not a prediction.
A dash, not a zero
If a run cannot produce a usable reading, we show an em dash. Zero means we asked and you were nowhere; a dash means the measurement failed. Printing one as the other invents a result we never observed.
See it on your own business
Around 60 seconds, no card, and every number links back to the answer it came from.