01
We run category prompts, not brand prompts
Each day we ask Doubao 豆包, DeepSeek, Kimi, Qwen 通义千问 and Yuanbao 腾讯元宝 the questions real shoppers ask — "best skincare brands", "affordable running shoes" — with no brand named in the prompt.
China GEO · Category leaderboard
Who’s winning AI search in your category?
Every day we ask China’s five leading AI models what to buy — then publish the brands they name. Pick your category and see who gets recommended.
0 models
Chinese AI models tracked — Doubao, DeepSeek, Kimi, Qwen, Yuanbao
0 categories
Categories covered at launch
0 dimensions
GEO Score dimensions — Presence, Ranking, Perception, Persuasion, Consistency
Daily
Prompts re-run and re-scored across every category
How it works
From prompt to public ranking.
This measures how AI assistants describe and recommend brands. It does not reflect product quality or brand reputation.
The leaderboard
Brand rankings by category.
Ranked by GEO Score across all five tracked models.
Accessories · 配饰
Data from 11 October 2026 · tracked across 5 AI models
Score = Presence + Ranking + Perception + Persuasion + Consistency
| # | Brand | Score | Visibility | Avg. position |
|---|---|---|---|---|
1New | HE Hermès | 83 | 20% | 2.6 |
2New | CH Chow Tai Fook | 79 | 31% | 3.4 |
3New | AP APM Monaco | 79 | 41% | 3.8 |
4New | CA Cartier | 77 | 18% | 3.0 |
5New | PA Pandora | 77 | 39% | 3.4 |
6New | 周大 周大福 | 76 | 11% | 3.2 |
7New | SW Swarovski | 72 | 39% | 3.3 |
8New | CH Chanel | 72 | 17% | 3.5 |
9New | TI Tiffany & Co. | 71 | 18% | 4.0 |
10New | AJ AJIDOU | 71 | 28% | 3.0 |
A sampled, directional signal — not a certified market-share measurement.Check your brand’s full ranking
Quick answers
Common questions.
We run a fixed set of category-level prompts — not brand-specific ones — through Doubao 豆包, DeepSeek, Kimi, Qwen 通义千问 and Yuanbao 腾讯元宝, and record every brand mentioned in the responses along with its position and sentiment.
Prompts are re-run and re-scored daily. The board shows the most recent completed day, which is displayed above the table.
If the tracked models didn't mention your brand for that category on that day, it won't appear. That absence is itself a signal — book a free consultation to see your brand's full position and what's driving it.
This measures how generative AI assistants describe and recommend brands in conversational answers — a distinct and fast-growing discovery channel, separate from traditional search or marketplace rankings.
Yes — get in touch if you believe your brand's data is inaccurate and we'll review it.
Want the full picture?
Track your brand across every model.
Track your brand and your named competitors across all five dimensions and all five AI models, with reporting that goes well beyond this public board.