King of AEO UK
How the UK figures on this site were actually measured
Every UK figure quoted on this site comes from one source: the round 0 baseline, measured on 4 October 2026 using six frozen prompts run with UK locale settings. This page sets out exactly how that UK slice was built, which engines contributed to it, which did not, and why a UK reader should treat six answers per engine as a small, early sample rather than a settled picture.
What actually makes an AI answer count as 'UK' in this data?
Each of the six frozen prompts was sent again with United Kingdom locale and language settings, through DataForSEO's search-engine simulation where relevant, or through each engine's own UK-region access point. The ten Google results checked for the UK slice were likewise pulled from UK-targeted search, not a generic or US-default view. Nothing in the UK figures on this site comes from a US or Australian run relabelled as UK.
Part 2
Which of the nine round 0 sources actually produced a UK-specific sample?
Six did: ChatGPT via DataForSEO, ChatGPT accessed logged out, Gemini via DataForSEO, the Gemini API, Google AI Mode and Google AI Overview. Three did not produce a market-specific UK row at all. Claude, queried through CLI web search, was run once without a market split. Perplexity, queried logged in, was only run for Australia in round 0; no UK or US Perplexity sample exists yet. A UK figure that looks lower than expected for Perplexity is not a real result; it is simply a source that has not been tested for this market yet.
Part 3
Why did the Gemini API return fewer UK answers than the other engines?
Five of the six UK-sliced sources returned all six prompts cleanly. The Gemini API returned only four of the six UK prompts in round 0, the one incomplete run in the UK sample. That shortfall is why the UK total sits at 44 answers rather than 46, and it is recorded as a known gap in the underlying data rather than smoothed over or backfilled with an assumed result.
Part 4
Why should six answers per engine be read with caution?
Six frozen prompts per engine, per market, is enough to spot a pattern worth measuring again, not enough to call a settled result. A single additional answer naming a different person first can move a UK percentage by more than fifteen points on a base that small. That is exactly why round 0 carried no title, why this site labels every UK figure as reference-only, and why the method matters as much as the headline number. The next UK-relevant measurement is round 1, on 13 October 2026, published 14 October.
Part 5
Will the UK method change for round 1?
The published plan keeps the same frozen-prompt approach for round 1: UK locale settings, the same set of AI sources where each one can be reached, and a UK-targeted Google check alongside it. Round 1 also folds the AI visibility score into the wider 100-point framework, where it contributes up to 48 of the available points rather than standing alone as it did in round 0. Whether the Gemini API returns a full six UK prompts next time, closing the one gap described above, will be visible once round 1's underlying data is published alongside the 14 October decision.
Check the full leaderboard
This site covers the UK slice. The official leaderboard combines all three markets and is the only place scores and the title are decided.