Korea AI Pulse
한국어
About

A new brand ranking,
for an era where AI answers.

A growing number of people in South Korea now ask GPT, Claude, Gemini and Grok— rather than Naver, the country’s dominant search portal — things like “which vet clinic is good at patellar luxation surgery”, “which cancer insurance is worth buying”, “which fintech is a good place to work”. But nobody was keeping a record of why a brand appears in those answers, or why it doesn’t. Korea AI Pulse records that by measuring, not by guessing.

Tracked questions
81
Verticals
7
Collections to date
714
Citations captured
1,371

What we do

  1. 1. Define 81 questions drawn from what people actually type into AI — vet clinics 22, law firms 13, restaurants 14, tech and fintech 13, insurance 9, fortune telling 5, GEO itself 5.
  2. 2. Every week, put the same questions to GPT, Claude, Gemini and Grok — in Korean, with no prompt engineering, exactly as a user would phrase them.
  3. 3. Archive the brand names that appeared, the order they appeared in, the stated reasoning and the source URLs, verbatim. Answers that named nobody are recorded too, as open spots.
  4. 4. Publish it as statistics (engine behaviour, open spots), source analysis (566 cited domains, ranked) and the collection log (every raw answer).
  5. 5. If you want a report for a specific brand, ask on contact — it’s free.

Why GEO for the Korean market specifically

Consumer research is moving out of the search box and into chat interfaces. GEO (Generative Engine Optimization)is how you understand and respond to visibility inside that new surface — the AI’s answer. Global GEO tools are built around English, so they handle Korean-language questions, Korean place names and citations to platforms like Naver blogs, Lawtalk and JobPlanet poorly or not at all. Korea AI Pulse builds measured data specifically around the Korean language, Korean domains and Korean brands.

What the data says

These aren’t marketing lines — they are computed from 714 collected records and change as more data lands.

  1. 1. One engine searches, another answers from memory — Grok runs a real web search on 100% of its answers; Claude manages only 72%. When an engine doesn't search, only brands already baked into its training data can appear.
  2. 2. Korea's local web belongs to Grok — Grok produces 47% of all citations to Korean local sources — Naver blogs, Instagram, community forums. The other engines stick to official sites and platform profiles, so local review marketing effectively reaches one engine out of four.
  3. 3. The engine that names the most businesses — Gemini names an average of 5.9 businesses or services per answer. It is the engine where competition for a named slot is fiercest.
  4. 4. The most open category — Law firms — Across the four engines, Law firms questions have the lowest rate of naming anyone at all, at 75%. Nobody has locked this category down yet, which makes it the easiest to enter.
  5. 5. The engines get things wrong.In our data we found a clinic in Suwon filed as a Seoul clinic, and a real company’s product attributed to a different company entirely. Which is exactly why how your brand is described in AI answers is worth checking regularly.

A note for brand owners

The clinic, firm, restaurant, company and product names shown on this site are the strings each engine returned, reproduced as-is and left in Korean. This is factual archiving, not promotion, and per-brand reports are provided free. If something needs correcting, or you’d like your own report, message us on contact.

Read next