The Pitot Index® is a standing record of how leading artificial-intelligence systems evaluate and recommend brands. Every measured brand is put through the same battery of questions, under the same controlled conditions, across nine leading AI models — and the readings are published as a single comparable baseline.

Through controlled measurement, not observation. Rather than sampling what people happen to ask AI, ARES-C — our measurement engine — applies a standardised evaluation framework to every brand and every model, cycle after cycle. The rules of the test never change between brands; only the brand and the model do.

The full approach, including how cultural variance across model cohorts is handled, is set out on the Methodology page.

Yes. The published record — current standings, weekly movement and the news wire — is free, and it is refreshed every cycle. The paid quarterly reports sit on top of that weekly record, with the depth the public record does not carry: full rankings, attribute-level readings, model-by-model and cultural breakdowns, and quarter-on-quarter analysis.

Measurement is weekly. ARES-C closes one benchmark cycle every week, on a fixed schedule, and the published record is refreshed as each cycle completes. The deep-dive reports are published quarterly and cover the full quarter’s cycles: quarterly is the reporting cadence, not the measurement cadence. The publication date of the next report is shown on the Reports page.

The Index is built and operated by Pitot Group Ltd, a London-based strategic marketing and AI company that builds its own measurement systems. PITOT INDEX® is a registered trade mark, and the measurement engine behind it — ARES-C — is proprietary to Pitot Group.

A score is an automated reading of what third-party AI models output about a brand under our standardised test — not our opinion of the brand, and not a statement about its actual quality or conduct. It tells you how AI, in aggregate, evaluates and recommends that brand at a point in time. Scores roll up into rating bands so that positions can be compared at a glance.

Because the models moved. AI systems change as they are retrained, updated and re-grounded on the live web, so a brand’s reading can shift without the brand doing anything at all. That is precisely what the Index exists to capture: the methodology stays fixed between measurement periods, so movement in the record reflects movement in AI — not a change in the rules of the test.

Neither. A model carries none of our human filters — no loyalty, no nostalgia, no preference. The Index does not report how people see a brand; it reports how AI rates it under identical conditions. When the two disagree, that distance is the information: it is often the first sign of a gap between a brand’s self-image and how machines are learning to describe it.

No. Pitot does not access, purchase or analyse private user conversations with AI systems, and the Index requires no personal data. Every prompt used in measurement is created for the research process and is independent of any individual user — the Index measures the behaviour of AI systems, not the behaviour of the people using them.

No. A brand’s position is determined by the ARES-C measurement process alone. Commercial relationships, report purchases and participation in Pitot services have no influence on scores or rankings — paid access buys deeper intelligence, never a different measurement.

Each quarterly report covers a full industry across every weekly cycle in the quarter: complete rankings and scores, attribute-level readings, model-by-model and cultural-register breakdowns, movement against previous quarters, and the analysis behind them. Reports are delivered as watermarked PDFs within one business day. Coverage runs from a single industry to a cluster to the full Index — see Pricing.

Every report is sold in three licence tiers: Individual (one named reader), Team (up to five named readers) and Enterprise (up to fifteen named readers, plus the right to quote excerpts and scores externally with attribution). All tiers are for internal use except where the Enterprise licence says otherwise — the detail lives in the Terms & Conditions.

Reports are sold to businesses only and supplied as soon as payment is confirmed. Because a report’s entire value is the information inside it, all sales are final once a report has been downloaded or accessed; a corrupted, incomplete or materially wrong file is replaced, or refunded if we cannot replace it. Subscriptions can be cancelled online at any time, effective from the end of the current billing period.

Brief quotation of individual published scores for criticism, review or news reporting is welcome, with attribution to “Pitot Index®”. Systematic republication, scraping, and any use of Index data to train, calibrate or validate AI models or competing datasets is not permitted. Enterprise licence holders may quote report excerpts and scores in their external materials with the attribution format set out in the Terms.

Pulse is a quick AI brand health-check. We name your brand to the leading AI systems, examine what they say about it under a controlled, versioned method, and deliver the result as a report the same business day.

It is the first step: one snapshot of where your brand stands in AI, and whether anything needs attention.

The Pitot Index is unprompted: the AI systems are never told which brands to discuss, and the Index records which brands they surface and how they rank them, every week. Pulse is prompted: your brand is named and examined directly.

The two answer different questions, so the same brand can come out very differently on each. A Pulse result is never comparable with an Index rating, and buying a Pulse has no effect on the Index.

A PDF report of about ten pages, in two editions: one for reading on screen and a white edition for printing. It contains your overall result on a five-step scale (Dark, Faint, Uneven, Established, Strong), your result on each of the four pillars, how closely what AI returns matches the positioning you gave us, the three most likely causes with where to start on each, and a page recording the method and the model versions that answered.

You can download a specimen report for a fictional brand.

Orders paid by 14:00 UK time on a business day are delivered by 18:00 the same day. Later orders, and orders placed at weekends or on UK bank holidays, are delivered by 10:00 the next business morning. Your delivery time is shown before you pay and on your confirmation.

You choose by sphere. Western Base uses three leading Western AI assistants; All models adds three more, including one that searches the live web. The Eastern sphere uses leading Chinese AI models, and the Middle East & Africa sphere uses Arabic-first AI built in the region. All AI systems runs every sphere in one examination.

Every system and the exact version that answered are listed on the report’s Method & versions page.

No. Measurement and scoring are fully automated under the published Pulse Method. Before a report is issued it is checked against a fixed list (the right brand, a complete version record, nothing withheld by the method disclosed). The results themselves are never adjusted.

It can be either. A single Pulse is a one-off snapshot, so it cannot show a trend. The monthly Pulse runs the same check every month, under the same method, option and market, and each report shows how your result has moved since your first check. Monthly costs about 30% less per check, with a minimum term of six months.

You pay by card through Stripe; Apple Pay and Google Pay appear where your device supports them. Prices exclude VAT. UK buyers pay VAT at 20%; business buyers elsewhere who give a valid VAT number are not charged UK VAT (reverse charge). A receipt and an invoice are emailed to you automatically.

You can raise a factual or methodological concern at any time, and we correct demonstrable errors in the application of our own methodology. What we do not do is alter scores on request: a reading is a reading. If a model is describing your brand in a way you did not expect, that is a finding to act on — not an error in the record.

The AI decides. Brand selection is part of the test sequence itself: within each industry, the measured models are asked which brands lead, and the top ten they converge on becomes the measured field. Nobody at Pitot places a brand in the Index, and brands cannot buy their way in or out — when AI starts naming a brand among an industry's leaders, the Index follows. If you want to understand how AI currently reads your own brand, the fastest route is a brand audit with Pitot Group.