YASO

Case study · China GEO

Borghese: 29 places up in China's AI search in one week.

How China GEO increased Borghese's AI Search Score. Our tool found where AI was missing the brand, filled the gap, and moved Borghese's ranking from #61 to #32 in a single test week.

Key results

+29 rank

AI Search Score, from #61 to #32

3/3

Target AI models increased visibility

+42.9%

Share of voice

5.4×

Target source citations

Borghese white mud mask

Background

AI search changes
how brands engage with customers.

Borghese is the Italian skincare house founded in Rome in 1957, famous for its royal heritage and luxury line of activated mud masks. In the last year, working with YASO, the brand has seen a +154% increase in net revenue across social commerce, driven by a disciplined content-to-conversion strategy led by our ecommerce team.

But the ecommerce landscape in China has seen a seismic shift with the launch and sharp adoption of AI assistants. A growing share of Chinese consumers no longer start their buying journey in a search bar or scrolling on RedNote. They ask an AI assistant.

More than 700 million people in China now use generative AI tools¹, and around 70% of Chinese consumers use it for shopping search, visual search and customer service². AI search now sits at the search-and-decide step, the last moment before checkout. If your brand is not in that answer, you are not in the decision.

1. China Internet Network Information Center (CNNIC), September 2026.

2. 2026 Global AI Agent Market Research Report, Frost & Sullivan, 2026.

Young Chinese consumer holding a smartphone

Test breakdown

China GEO's action plan in practice

How China GEO made an Italian luxury brand show up in China's AI models.

The AI Search Score is what allows brands to track and diagnose their performance over time. China GEO's ability to generate prioritised action plans and content output based on this diagnostic work is what allows brands to move the needle.

We isolated the test period. No other marketing activities took place during this time. We followed the exact strategic recommendations laid out for our team by China GEO and recorded the results after just a 5-working day window.

Here's what the tool was capable of.

01 Baseline measurement

China GEO's baseline analysis

Every China GEO report starts with a radar. Five measures, each scored out of 20, add up to your AI Search Score out of 100. The shape tells you more than the number: it tells you exactly where your brand is losing ground to competitor brands. Borghese was scoring 57/100 in the period prior to our test.

01 / 06

Results

What changed in one week

We measured the test week against the baseline week.
Here are the results that stood out from our analysis.

Overall AI search ranking

+29 places

Rose from Ranked #61 to Ranked #32.

Overall AI search ranking

Brand ranking

#BrandMentionsScore
1124284.0
283280.4
37473.7
…
32Borghese31061.2

Topic-specific AI search ranking

Increased

Ranking for 3/5 key topics

tracked for the brand

Topic-specific AI search ranking

Topic rankings

TopicRanking beforeRanking after
Skin Concerns#36#15
Where to Buy#38#27
Buying Guide#16#7

China GEO gives brands the ability to track their AI ranking in specific customer search scenarios. For Borghese, the tool was able to increase ranking in 3/5 key scenarios.

Source gaps closed

2 of 2

SMZDM.com's gap reduced 93%

Zhihu.com's gap dropped off the list

Source gaps closed

Source gaps

#DomainSource typeGap score
1zhihu.comSocial Media220.3 → 0
5smzdm.comIndustry Review & Blog122.1 → 9.1

The diagnosis found the sources AI trusted but where Borghese was missing. A week after publishing with China GEO, the SMZDM.com gap had been reduced 92.5%. Zhihu.com, the biggest gap, had closed entirely.

Citations from target sources

Increased

SMZDM.com increased 3-fold

Zhihu.com increased 5-fold

Citations from target sources

Domain detail

#Source typeDomainCitations
1Social Mediazhihu.com28 → 151
2Industry Review & Blogsmzdm.com24 → 72

A closed gap only counts if AI uses the content. Across the two target platforms, AI assistants cited Borghese 223 times in test week, up from 52. AI was now drawing on Borghese content from the sources it already trusted.

Target AI model visibility

Increased

Qwen +79%

DeepSeek +48%

Yuanbao +16%

Target AI model visibility

Change in visibility score

Qwen
+78.8%
DeepSeek
+48.2%
Yuanbao
+16.0%
Kimi
−2.6%
Doubao
−3.6%

Before our work, the diagnosis showed three specific models pulling Borghese's score down. Within a week, all three showed increased visibility led by Qwen, Alibaba's AI assistant. Models which were not targets in this test showed minimal deviation.

Overall AI visibility and share of voice

Increased

AI visibility +24%

Share of voice +43%

Overall AI visibility and share of voice

Change week on week

Share of voice
+42.9%
AI visibility
+23.7%

Visibility, the share of relevant AI answers that mention Borghese, increased with the brand now appearing in one in six search results. Share of voice is Borghese's share of all brand mentions against competitors, and rose by more than 40% in the same week.

What we learned

Not a one-off project.
A repeatable growth loop.

Borghese's test shows what China GEO can do before a brand commits to a full programme.

01 · Measure

Results in as little as one working week

China GEO can move your brand's AI search performance within a single working week. Borghese climbed 29 places in one Monday-to-Friday test.

02 · Diagnose

A precise diagnosis, not a guess

China GEO's proprietary AI Search Score breaks your performance into five measures across five AI models, so you see exactly where you are losing ground. For Borghese, it traced a 61st-place ranking back to one weak measure and three AI models.

03 · Recommend

Every action ranked by impact

Recommendations turn the diagnosis into a ranked action list: what to create, on which platform, for which AI models. Your team starts with the work that moves the score most, not the work that is easiest to brief.

04 · Create

More impact, less workload

China GEO turns each recommendation into a brief and a first draft in minutes, in English or Mandarin. Specialists spend their time on judgement, cultural nuance and getting the brand right.

05 · Publish

Where you publish matters more than how much

Borghese did not need more general content. It needed to be present in the specific sources AI already trusts. Content on two platforms took citations from 52 to 223, and lifted visibility on all three target AI models.

How China GEO
works for your team

01

Measure

02

Diagnose

03

Recommend

04

Create

05

Publish

Each cycle starts from the last one's reading. Borghese's next cycle starts with persuasion: getting AI to recommend the brand, not just cite it. Continuous tracking is what turns one result into a compounding one.

Plug in. Scale up.

Get your brand on China GEO today.

What does your radar say? Get your free AI visibility report, or book a 30-minute demo and we will walk you through your brand's score, source gaps and first recommendations.