The AI Desire Index
AI assistants are a second kind of consumer, and they have views of their own. Every month 7 AI models are asked how well they know the top 250 brands in each of 25 markets, and how far they agree with the 36 statements of the Havas Science of Desire 2.0: how much each brand is wanted, belonged to and returned to. Each brand is scored out of 10 and indexed against the other brands in its segment.
6,239 brands · 25 markets · 7 AI models · 43,689 model ratings of a brand
Every market
The three highest-Desire brands in each market, across all of its segments (Desire out of 10). Each market's page ranks every segment and indexes each brand within it.
North America
| Market | Most desired brands | Brands | Month |
|---|---|---|---|
| United States | Minecraft 7.8 · Nike 7.7 · Netflix 7.5 | 250 | November 2026 |
| Canada | YouTube 7.9 · Minecraft 7.8 · Nike 7.6 | 250 | November 2026 |
Europe
| Market | Most desired brands | Brands | Month |
|---|---|---|---|
| United Kingdom | Nike 7.5 · Spotify 7.5 · Nintendo 7.5 | 250 | November 2026 |
| Germany | YouTube 7.5 · Minecraft 7.5 · Spotify 7.4 | 250 | November 2026 |
| France | YouTube 7.8 · Steam 7.6 · Nintendo 7.4 | 250 | November 2026 |
| Spain | Spotify 7.9 · Steam 7.8 · Nike 7.6 | 250 | November 2026 |
| Italy | Nike 7.8 · Spotify 7.7 · YouTube 7.6 | 250 | November 2026 |
| Netherlands | YouTube 7.9 · Spotify 7.8 · Steam 7.4 | 250 | November 2026 |
| Poland | YouTube 7.8 · Steam 7.6 · TikTok 7.5 | 250 | November 2026 |
| Turkey | Nike 7.8 · YouTube 7.8 · Spotify 7.7 | 250 | November 2026 |
Latin America
| Market | Most desired brands | Brands | Month |
|---|---|---|---|
| Brazil | YouTube 8.4 · Pix 8.1 · Spotify 7.8 | 248 | November 2026 |
| Mexico | Minecraft 7.8 · Nike 7.7 · YouTube 7.7 | 250 | November 2026 |
| Argentina | Spotify 7.8 · YouTube 7.8 · Mercado Pago 7.6 | 249 | November 2026 |
| Colombia | YouTube 8.1 · Minecraft 8.0 · Spotify 7.9 | 250 | November 2026 |
| Chile | YouTube 7.8 · Spotify 7.7 · Nike 7.6 | 249 | November 2026 |
Asia-Pacific
| Market | Most desired brands | Brands | Month |
|---|---|---|---|
| China | WeChat 8.1 · WeChat Pay 7.8 · Huawei 7.6 | 249 | November 2026 |
| Japan | Nintendo 7.9 · LINE 7.7 · Apple 7.5 | 250 | November 2026 |
| India | YouTube 8.6 · Google 7.8 · WhatsApp 7.7 | 249 | November 2026 |
| South Korea | YouTube 8.0 · Toss 7.8 · Nike 7.8 | 249 | November 2026 |
| Australia | YouTube 7.7 · Minecraft 7.7 · Canva 7.6 | 250 | November 2026 |
| Indonesia | Nike 8.0 · YouTube 7.9 · TikTok 7.8 | 250 | November 2026 |
| Singapore | YouTube 7.7 · Apple Store 7.5 · Nike 7.5 | 249 | November 2026 |
Middle East & Africa
| Market | Most desired brands | Brands | Month |
|---|---|---|---|
| United Arab Emirates | YouTube 7.9 · Nike 7.9 · Spotify 7.7 | 250 | November 2026 |
| Saudi Arabia | YouTube 8.3 · Nike 7.9 · Apple 7.7 | 248 | November 2026 |
| South Africa | Nike 7.7 · YouTube 7.7 · Spotify 7.7 | 249 | November 2026 |
Method & limitations
Once a month the top 250 brands in each of 25 markets are put to 7 AI models (Claude, Perplexity, Gemini, ChatGPT, Grok, DeepSeek and Meta AI), the way a survey asks a respondent: first how well the model knows each brand, then how far it agrees with each of the 36 statements of the Havas Science of Desire 2.0, on a 1–7 scale. A model that rates its familiarity with a brand below 3 is left out for that brand. Each statement is the mean of the remaining models' ratings; statements average into fifteen dimensions, dimensions into the three Drivers of Desire, and the drivers into Desire, each out of 10. Each brand's index is its Desire divided by the average Desire of the brands shown in its segment in that market, × 100: 100 is the segment average, 115 is fifteen per cent above it. Change is the move in index points since the market's previous month.
A brand is shown only when at least 4 models knew it well enough to score it, and a segment only when at least 3 brands are shown. Only finished monthly runs are published. These are the AI models' own ratings when asked, not a survey of people and not an endorsement; brands are compared within a segment and market, not across them. See the full methodology.
Where next
This board measures what AI wants from a brand. These answer what comes next — which brands AI names, which websites feed those answers, and how any of it is counted.