From snapshot to baseline: why AI brand measurement has to be repeated

One reading — whether a Pitot Index® week, a Pulse or an Audit — is a snapshot. It tells you where a brand stands under that instrument at that moment. A baseline is different: it comes from repeating the same standardised measurement over time, so later readings can be set beside earlier ones.

The published Index covers the top ten brands in each industry, and that public weekly record is the industry-wide AI Brand Rating Baseline. You do not need to be in the top ten to have an AI Brand Rating baseline. Pulse runs a standardised method, and any change to it is versioned and dated, so any brand can build its own baseline by repeating Pulse. Each run is a snapshot; the repeats together become a baseline set for that brand. That set is not an Index ranking, and it is never compared with the unprompted Index line.

AI Brand Understanding (AIBU) is the category this belongs to: how and why AI models understand a brand. AI Brand Understanding is measured as an AI Brand Rating. How the Index, Pulse and Audit are built is set out in our measurement explainer; what matters here is that a single reading is not yet a baseline.

Like for like only

Each instrument gives its own AI Brand Rating reading, comparable only with itself. Read Index with Index, under the same method version. Read a Pulse against earlier Pulses. The Pitot Index is unprompted; Pulse and the Audit are prompted — the brand is named. Prompted and unprompted readings answer different questions and are never compared.

What a brand baseline lets you see

GEO and AEO are how teams work on where a brand appears in AI answers — mentions, share of voice, sentiment, citations. That work can be read more clearly when it is set against a stable AI Brand Rating baseline. The baseline does not replace GEO or AEO. In our view, a change in mentions without a change in the rating, or a change in the rating without a change in mentions, is information either way — but only if the rating sits on a baseline built from repeated, comparable readings.

The same applies, in our view, to corrections and other brand activity: they can be read against the baseline to see whether how AI understands the brand moved.

The weekly Index shows where brands in the published top ten stand and where the gaps are. The quarterly industry reports go deeper across an industry — full rankings, brand-by-brand scores and analysis, and how the readings diverge by region — on the weekly record. Pulse sets out where one brand stands and the most likely causes. The Audit confirms the causes and specifies the corrections. Repeating Pulse works like a periodic check-up: take one health check, or one every month, with the trend since your first. Each check tells you how things stand; check after check shows whether they are improving or deteriorating. For in-house teams doing GEO and AEO work, in our view, that record can show which gaps to focus on and how the rating moved while the work ran.

What a baseline is not

An AI Brand Rating baseline is an automated reading of what AI models output under a standardised test, built up over repeated runs. It is not a measure of what buyers think, how often real users are shown a brand, or how often a brand is recommended in everyday use. It is not a statement about a brand's quality or conduct. Search rankings, AI mentions, cited sources and sales remain separate measures.

No measurement can promise a place in an AI answer. What a baseline can do is give a reference so that claims of improvement, decline or drift have repeated, comparable readings to stand on.

The machines are forming a reading whether a brand watches or not. The baseline is how you follow that reading, consistently, over time.