Guide · 3 min read
What is AI visibility?
A plain explanation of the single number at the centre of the Index, and how to read it without overreading it.
A share of answers, not a share of searches
When someone asks an AI assistant which bank to open an account with or which broadband provider to choose, the reply usually names a handful of brands. AI visibility, as the UK AI Visibility Index uses the term, is the proportion of measured answers in a category that name a particular brand at least once.
If a category has 1,200 completed answers in its rolling window and a brand appears in 300 of them, its visibility is 25%. An answer counts once however many times it repeats the name. That keeps the figure tied to a simple question a reader can check: in how many answers did this brand come up at all?
Visibility is not a market share, a sales estimate or a measure of traffic. It does not tell you how many people asked a similar question in real life, and it does not say whether the brand was praised or criticised. It records presence in the answer.
Where the answers come from
Each category has a fixed panel of naturally phrased UK questions. None of them contains a brand name. The same panel is sent every week to each measured assistant through its public API, using the default consumer model with web search switched on. The methodology page lists the current models and the dates each was adopted.
Brands are not chosen in advance. They are detected in the text of the answers, merged where reviewed name variants clearly refer to the same company, and only then counted. A brand is in the Index because an assistant named it.
Because generative models give different replies to the same question from one run to the next, a single week is noisy. The published figure therefore uses a rolling window of several weeks. The window length for each category is shown on its page, next to the number of answers it contains.
Reading the number sensibly
Visibility is most useful as a comparison within one category. A brand on 40% in banking and another on 12% in the same category were named at very different rates by the same panel of questions in the same weeks. Comparing a banking figure with a furniture figure says much less, because the panels, the number of plausible brands and the assistants' habits all differ.
Small differences deserve caution. Two brands a few tenths of a point apart may swap places next week without anything meaningful having changed. The answer count beside every figure is there so you can judge how much evidence sits behind it.
It is also worth reading the per-assistant figures. A brand can have solid overall visibility because one assistant names it in most answers while another rarely mentions it. The overall number is an average over the measured assistants' answers, so it can hide that kind of split.
What visibility cannot tell you
The Index samples a bounded panel of questions. Real people phrase things in countless other ways, add personal context and ask follow-up questions. Visibility describes what happened for this panel, in this locale, with these models, during this window. It is a consistent reference point, not a complete picture of every conversation.
Nor does the Index claim to know why an assistant named a brand. The assistants do not publish their internal ranking logic, and the Index does not attempt to reverse-engineer it. What can be shown is the observable output: which brands appeared, how often, in which position, and which sources were cited alongside them.
Used with those limits in mind, the figure is a practical starting point. It lets a brand see whether it is part of the conversation in its category, how that compares with named competitors, and whether the position is moving over time.
Live evidence · In the week to 13 November 2026 (week 47, published 16 November)
Current category leaders
The brand with the highest visibility in each published category this week, with the answer count behind the figure.
| Category | Leading brand | Visibility | Answers |
|---|---|---|---|
| Banking and current accounts | NatWest | 43.3% | 900 |
| Investing and wealth | AJ Bell | 52.9% | 900 |
| Mortgages and personal finance | HSBC | 24.6% | 900 |
| Business finance and payments | Xero | 25.9% | 900 |
| Cars and motoring | Carwow | 37.0% | 900 |
| Fashion and clothing | Marks & Spencer | 43.1% | 900 |
Full definitions and limitations are in the methodology.