Does AI rate the same brand differently in different industries?

Yes, and sometimes by a wide margin. In the Pitot Index for cycle 2026-W41, published 4 October 2026, Mercedes-Benz scored 80.1 in Luxury Automotive and 66.0 in EV Automotive: 14.1 points apart in the same week. BMW (77.3 against 66.2) and Audi (72.9 against 63.4) showed the same pattern. Brands rated in two closely related industries read almost identically: Porsche scored 81.8 in both Luxury and Performance Automotive.

Twelve brands appear in the top ten of more than one Pitot Index industry in this cycle. Each industry is measured separately, with one standardised, unprompted test: the AI models are never told which brands to discuss, and each brand is rated in the company of the peers the models name for that industry. Reading those twelve side by side shows how much the industry frame changes the reading.

The figures

Pitot Index, cycle 2026-W41. Scores 0–100. Where a brand is rated in three industries, the table shows its highest and lowest. Mercedes-Benz also scores 74.9 (rank 3), BMW 72.6 (rank 4) and Audi 69.1 (rank 6) in Mainstream Automotive.

Three patterns

The German premium marques read lowest in EV Automotive. Mercedes-Benz, BMW and Audi are each rated in three industries, and each scores highest in Luxury Automotive and lowest in EV Automotive, with Mainstream Automotive in between. In EV Automotive, the top two are BYD (76.3) and Tesla (74.2), followed by Hyundai (69.2) and Kia (67.0). BMW is fifth, Mercedes-Benz sixth and Audi eighth.

Fashion houses read higher in fashion than in beauty. Chanel scores 83.7 in Luxury Fashion and 74.1 in Beauty, second in both. Dior scores 81.1 and 72.9, third and fifth. Beauty is closely packed at the top: Shiseido is first on 74.7, with Chanel and Estée Lauder on 74.1 and L'Oréal on 73.9, so the top four sit within 0.8 of a point.

Closely related industries read alike. The four marques rated in both Luxury and Performance Automotive differ by 1.8 points or less: Ferrari 1.8, Bugatti 1.3, Lamborghini 1.1 and Porsche 0.0.

How much of the gap is the industry?

Part of each gap belongs to the industry, not the brand. The ten brands in each industry's top ten have different averages. In 2026-W41, the averages of the ten published scores are 77.5 in Luxury Automotive, 70.8 in Mainstream Automotive and 66.9 in EV Automotive, and 77.8 in Luxury Fashion and 70.6 in Beauty. These match the category averages on Pitot Group's sector pages.

So the useful comparison is a brand against its own industry. On that basis:

• Mercedes-Benz is 2.6 points above the Luxury Automotive average and 4.1 above the Mainstream Automotive average, but 0.9 below the EV Automotive average. Of the three German marques, it changes most against its industry between Luxury and EV.
• BMW is close to the average in Luxury (0.2 below) and EV Automotive (0.7 below), and 1.8 above it in Mainstream.
• Audi is below the average in all three: 4.6 points in Luxury, 1.7 in Mainstream and 3.5 in EV.
• Chanel and Dior are above the average in both their industries: Chanel by 5.9 in Luxury Fashion and 3.5 in Beauty, Dior by 3.3 and 2.3.

Hyundai is the counter-example. Its score is almost the same in EV and Mainstream Automotive (69.2 against 69.8), but it ranks third in EV against fifth in Mainstream.

What this does and does not show

It shows that a brand's AI reading is not a single property. The same name, in the same week, is rated differently depending on the industry and the peers it is read against. Context is one of the five ARES-C dimensions for this reason: where a brand surfaces, how widely, and what it is associated with beyond its own category.

The 2026-W41 figures show where the gaps are. The cause sits one layer down, and finding it is what AI Brand Understanding is for. The quarterly industry reports go to that layer across a whole industry, with brand-by-brand analysis behind each reading and how the readings diverge by region. A single cycle is not a trend, and EV Automotive is also the youngest series in the Index: five weekly cycles, from 2026-W37.

Nor is it a measure of buyers. These are automated readings of what AI models output under a standardised test, not our opinion of any brand and not a statement about its quality or conduct. They don't show what consumers think, how often real users are shown these brands, or how often they are recommended in everyday use.

For a brand that competes in more than one category, there is a separate reading in each one, and a strong reading in one says little about the others.