Finding 01 — The same brand, the same week, four different verdicts.
Thirteen brands, all well known in their Taiwanese categories: Gogoro, Whoscall, KKday, Pinkoi, Dcard, LINE Bank, Coupang Taiwan, Dent&Co, FunNow, Rainbow Six Siege, Love and Deepspace, Anden Hud and iNewMe. Every question was asked the way a Taiwanese buyer would ask it, in Chinese.
| Across all thirteen brands | ChatGPT | Google AIO | Claude | Perplexity |
|---|---|---|---|---|
| Recommended the local brand | 38% | 76% | 56% | 75% |
| Ranked it first | 17% | 47% | 36% | 41% |
The spread is not a rounding error. Take Whoscall, the anti-scam caller-ID app built by Taiwanese company Gogolook. Google AI Overviews recommends it on 5 of 6 questions and ranks it first on 4. Claude recommends it on 4 of 6, Perplexity on 5 of 6. On ChatGPT, same questions, same week: 0 of 6. A consumer asking one assistant which anti-scam app to install gets Taiwan's category leader; asking another, they never hear the name.
Finding 02 — The engines are not reading the same amount. They are not close.
We counted the unique sources each engine disclosed per answer. The difference is an order of magnitude.
| Per answer | ChatGPT | Google AIO | Claude | Perplexity |
|---|---|---|---|---|
| Median sources cited | 0.5 | 5 | 10 | 14 |
| Answers citing nothing | 50% | 9% | 3% | 0% |
Half of ChatGPT's answers to these Chinese buyer questions arrived with no sources attached — written from what the model already knew. Perplexity never once answered without citing something. This was not driven by one unusual brand: zero-source answers appeared for eleven of the thirteen brands we scanned.
Finding 03 — Retrieval is what decides whether the local brand shows up.
Split all 307 answers by whether the engine cited anything, and the gap is stark.
| Answer type | Named the local brand |
|---|---|
| Built from cited sources (n=259) | 68% |
| No sources cited (n=48) | 25% |
The same split holds inside individual engines, which is the part that matters. ChatGPT recommends the local brand in 51% of its sourced answers but only 26% of its unsourced ones. Google AI Overviews: 80% versus 29%. The engine is not changing its opinion of the brand — it is changing whether it goes and looks.
The clearest confirmation is the brand that never triggered a memory-only answer. Gogoro drew zero unsourced answers across all 24 checks — every engine went and looked, every time — and it scored 96% recommendation share with the top slot in 19 of 24. When the lookup always happens, the local brand always shows up.
Finding 04 — When the lookup happens, it may read your rival's website.
Retrieval is necessary, not sufficient — because the engine still has to find you in what it reads. KKday is the sharpest case in our set. It is recommended in 71% of checks, so visibility is not the problem. But it takes the first slot in 2 of 24, its rival Klook is named in 23 answers, and the single most-cited source across the whole category is klook.com — cited 15 times, against 2 for kkday.com. The engines are answering questions about this category by reading the competitor's site.
That is the second job, and it is the one a brand can act on this quarter: be present in the sources these answers are actually built from. Across our thirteen scans those were community threads, comparison articles and app-store pages far more often than official sites — and in two categories a rival's own domain out-cited the brand's.
Method, and what this does not show
Thirteen brands across ten categories, six buyer questions each, four engines, one pass per cell: 312 cells, 307 scored (5 lost to engine rate limits). Questions were generated from each brand's own site and market, then asked verbatim in Traditional Chinese. Each answer was scored by the same rubric: is the brand named, is it recommended, is it ranked first, and which sources were cited.
Four honest limits. One: we count the sources an engine discloses. An engine may retrieve without showing citations, so "no sources" means "no sources shown", and the comparison between engines is a comparison of disclosed behaviour. Two: Google AI Overviews was queried through a Taiwan-located search, which gives it a structural local advantage the chat engines do not have — some of its lead is a geography setting, not a preference. Three: these thirteen brands are not equally famous, and a single pass per cell cannot separate a genuine gap from the ~15% run-to-run drift we have measured in three previous studies. Four: Gemini was excluded — its grounded runs did not return within our time budget.
What we are confident in: the retrieval split holds across engines, across categories, and across eleven of thirteen brands. What we are not claiming: that any engine is biased against Taiwan. The measured behaviour is about how often an engine looks things up, and the consequence falls hardest on brands whose fame is regional.
Is your brand in the memory, or only in the lookups?
Paste your site — one real buyer question across five engines, free, no email. The full scan tracks it weekly, engine by engine, with the sources behind every answer.