Category study · Taiwan market

Which Taiwanese skincare brands does AI recommend?

Consumers are handing buying questions to AI - Taiwan surveys put generative-AI usage at 43% of adults, with 38% of shoppers consulting AI before purchases. We put six real skincare questions to five engines, 30 checks: 93% returned concrete brand shortlists, no brand owns the list, and the answers are fed by Taiwanese beauty media - not brand websites.

By Hsiang Lee · Ainswer · Oct 2026

Method: six questions Taiwanese consumers actually ask ("Which serum suits sensitive skin?", "Which Taiwanese skincare brands make good gifts for friends abroad?"), put simultaneously to ChatGPT, Gemini, Claude, Google AI and Perplexity - 30 checks, every mention and citation logged. Demand-side context with sources and math: TWNIC 2025 (43% of Taiwanese adults have used generative AI), Mastercard 2026 via CNA (38% consult AI when shopping), Rithum 2026 (34% of AI shoppers ask about beauty). Different survey populations - estimates, not measurements. What we measure precisely is the other side: what AI answers when asked.

1 - 93%: skincare is a hot zone

28 of 30 checks (93%) returned concrete brand shortlists - not skincare chit-chat, but named brands with reasons to buy. Among the eight categories we have measured, skincare ranks near the top for answer density. The buying-decision doorway has moved.

2 - The list has no owner

The study's most important finding. Where study-abroad consulting has Crimson named 15 times and meditation apps are locked up by Calm and Headspace, Taiwan's skincare mentions are highly fragmented: over 15 brands appeared (NARUKO, Greenvines, Neogence, My Beauty Diary, For Beloved One...), nearly all exactly once; even the steadiest performer, DR.WU, took only about half the checks (16/30).

In a category with an incumbent, newcomers must displace someone. Here, the first brand to work AI visibility systematically is defining the list itself.

3 - Who feeds the answers: local beauty media, not brand sites

SourceCitedType
Bella (bella.tw)7Beauty media
Dcard6Community reviews
cosme Taiwan6Review platform
vocus6Creator platform
Beauty3216Beauty media
Marie Claire TW · my-best · PTT · UDN3-4Media & community

Note the contrast with US research, where beauty citations concentrate on Reddit, YouTube and Sephora: Taiwan's answer corpus is local - Bella, Dcard, cosme, Beauty321. To enter Taiwanese consumers' AI answers, these are the doors. Brand websites barely register.

4 - Three moves for brands

1. Work the six doors AI actually reads: the table is the map - real coverage in Bella or Beauty321 and genuine Dcard discussion beat any homepage redesign.
2. Give your site citable facts: ingredient tables, skin-type matrices, honest comparisons - AI cites specifics, not brand copy.
3. Measure weekly and claim the list early: the no-owner window will not stay open. Weekly fixed-question rescans show your movement against competitors - the first systematic mover becomes the future list.

Want to know whether your brand is on the list today? Run a free scan - one real buyer question, live →