AI search visibility for UK brands and agencies: AI Overviews, pounds, and numbers with the sample size on them
AI search is the umbrella the trade press uses. Answer engine optimisation (AEO) and generative engine optimisation (GEO) are what practitioners call the work. ClarAI is the scoreboard underneath either: it asks the engines your buyers' questions, keeps every answer verbatim, and prints the sample size beside every number. Priced in dollars, and this page says so plainly.
What changed for UK search
Google's AI Mode began rolling out in the UK on 28 July 2025. Google's own UK announcement says it appears as a tab on the Search results page and in the Google app for Android and iOS. That is the date to give a client who asks how long this has been live here: over a year.
Search Console now carries a Generative AI performance report. Google's help page says it shows how many times links to your site were shown in a generative AI feature on Google Search, covering AI Overviews and AI Mode, grouped by pages, countries, dates and devices. It reports impressions only: no clicks, no click-through rate, no position and no queries. As of 31 August 2026, Google says the report has rolled out to all websites worldwide. Search Engine Journal reported on 31 August 2026 that the June test began with a select group of UK site owners; Google's own launch post did not load for us, so treat that detail as reported rather than confirmed.
What the report cannot tell you is which brands the answer named, which pages it cited, or how the answer read. That is the gap ClarAI fills: for Google AI Overviews, and for ChatGPT, Claude and Gemini. We wrote up what Google's own guidance says earns a place in these features, and what it says does not.
What UK buyers type, verbatim
Google's UK autocomplete completes the task four ways: how to get my business on chatgpt, how to get my business listed on chatgpt, how to get my business recommended on chatgpt, and how to register my business on chatgpt. (Our own fetch, 9 September 2026.)
The last one carries a misconception worth killing plainly: there is no registration form. ChatGPT has no business listing to claim, no directory to submit to, and no paid placement inside its answers. The engines answer from what they retrieve: the pages they can crawl and read, the third-party pages that describe your category, and the answer-first pages that state plainly who is for whom.
So "listed" and "recommended" mean the same thing: you appear when the engine's sources appear. The work is being on those sources and being clear on your own pages. The measurement is asking the engines the questions your buyers ask and counting how often you are named, with the count of runs shown beside the result. ClarAI does the measurement. The free grader checks whether your site is readable enough to be retrieved in the first place, which is readiness, not visibility.
The other UK strings are tool-shaped: ai visibility checker free, ai visibility tracker, ai visibility expert. The free checker funnels into the grader; the tracker is the product; and the experts are the UK agencies this next section is written for.
For UK agencies selling GEO
Search "best GEO agencies UK" and the lists that rank name agencies, not the tool behind them. Every agency on those lists sells a programme it has to report on monthly.
The scoreboard behind the service
- The monthly report needs a number the client cannot pick apart: which engines, which prompts, how many runs, and the answers themselves. A Head of Marketing at a D2C brand will ask for the n; a scoreboard that prints it survives the question.
- A workspace per client, kept separate and access-checked on every query, with an org rollup across the whole book so the number you brief on and the number you send are the same number.
- Per-client approval routing with named reviewers and deadlines, for anything going out under a client's name. Nothing publishes by itself.
- A white-label PDF of a client's real visibility data carrying your agency's name. The agency page lists what is shipped and what is still roadmap, kept separate on purpose.
Priced for the book, not per client
- Agency is $299 a month for the account, with unlimited brand projects, so a full analysis run to win a pitch costs nothing extra and displaces no paying client.
- Ten clients is $29.90 a client a month. That is division on our published price, not a case study, and we have no customer revenue to model anything else from.
- Up to 10 team seats, 500 analyses a month as a fair-use ceiling, the org rollup, per-client routing and the white-label PDF are the Agency gates today. Your logo on the PDF and client portal links are on the roadmap, not shipped.
- Pro at $99 is the self-serve entry for an agency under roughly fifteen clients: unlimited brands, three seats, API and MCP, but no client workspaces and no PDF.
Pricing, in US dollars, stated plainly
Three plans, monthly, no annual price yet. The full comparison is on the pricing page; these are the same numbers.
| Plan | Price | What it includes today |
|---|---|---|
| Free | $0. Seven days of the full product with no card, then a free Watch plan. | One brand project, one seat, 5 analyses and 10 simulations a month, the headline score with its trend and one opportunity a week. |
| Pro | $99 a month | Unlimited brand projects, up to 3 seats, 150 analyses and 200 simulations a month, the full per-engine and competitor detail, API and MCP access, chart PNG export. |
| Agency | $299 a month | Everything in Pro, up to 10 seats, 500 analyses and 500 simulations a month, a workspace per client, the org rollup, per-client approval routing, and the white-label PDF carrying your agency's name. |
What ClarAI monitors, and what it does not
The engines
- Four. ChatGPT, Claude and Gemini run through their official APIs, with Claude's answers grounded by web search, and Google AI Overviews is read from a live Google search.
- Google AI Mode and Perplexity are not monitored. Your account shows the live engine count.
- We measure through official engine APIs plus grounded search, and we say on the page that an API answer can differ from a logged-in session. We do not scrape a consumer surface and call it the same thing.
The numbers
- Every headline number carries its sample size: "named in 3 of 8 runs", "4 of 72 prompts measured". A prompt that was never measured is shown as unmeasured, never as absent.
- Every measured answer is kept verbatim with its citations, so the number can be audited by the person paying for it.
- After three runs a variance band appears, so a small move reads as a tie rather than a win.
- The free grader measures readiness, not visibility: a technical check on public signals, about twenty seconds, no signup. A grade-A site can still go unmentioned.
Questions UK teams ask us
Do you bill in pounds?
Is my data held in the UK?
Does ClarAI track Google AI Mode and AI Overviews?
Is there a UK-specific version of the product?
Why does every number carry a sample size?
Do you have UK customers I can talk to?
Sources
- Google, The Keyword UK: AI Mode now available on Google Search in the UK (28 July 2025) (read 9 September 2026)
- Search Console Help: Generative AI performance report (Search) (read 9 September 2026)
- Search Engine Journal: Google Search Console AI Reports Rolled Out Worldwide (31 August 2026) (read 9 September 2026; the only source for the UK test group detail)
- brightonSEO: Our next UK event, 8 to 9 October 2026 (read 9 September 2026)
- Google UK autocomplete strings for the ChatGPT task queries (ClarAI's own fetch, 9 September 2026)
- ClarAI plans, prices and plan gates (the shipped pricing data and billing rules, 9 September 2026)
Start with the domain in your next client deck.
Paste it into the grader: no signup, about twenty seconds, a readiness grade with the checks it ran. Then create an account and see what the engines actually say, with the sample size on it.