ANSWER ENGINE OPTIMIZATION
AEO in Korea —
measured, with the
conditions published
Navirang is an answer engine optimization agency working in Korean-language AI search. We record whether a brand is mentioned and cited across the answer engines that matter here, and we publish how we measured it.
SHORT ANSWER
Navirang is an answer engine optimization agency based in Korea. We measure whether a brand is mentioned and cited across 7 answer engines — including Naver AI Search, which no overseas agency covers — and publish our measurement conditions so the numbers can be checked.
THE BASICS
What Navirang is
Each row is written to stand on its own if it is quoted without the rest.
- What we do
- Answer engine optimization (AEO): making a brand findable, readable and citable by AI answer engines, and measuring whether that actually happens.
- Where we are
- NAVIRANG (주식회사 나비랑) operates from Korea. Our working market is Korean-language AI search, which behaves differently from English-language AI search.
- What we measure
- 7 answer engines: ChatGPT, Google AI Overviews, Perplexity, Claude, Naver AI Search, Copilot, Gemini. Mentions and citations are recorded separately, because they move independently.
- The denominator
- Our own baseline is 40 questions × 7 engines = 280 cells. A citation rate without a stated denominator cannot be verified, so we always state ours.
- What we do not promise
- Guaranteed placement. No engine publishes its selection criteria, and OpenAI states in its own documentation that there is no way to guarantee top placement.
- What is unusual here
- We publish the measurement protocol, the question set and the raw data for measurements of our own site — including the ones that went badly.
WHY KOREA IS DIFFERENT
Three things that surprise teams entering Korea
These are the items an international AI-search checklist does not cover.
- 01
Naver is a separate index, not a smaller Google
Naver AI Search draws heavily on Naver's own services. Its crawler is Yeti and its entry path is Naver Search Advisor — neither of which appears in a standard international SEO checklist.
- 02
Naver Blog blocks overseas AI crawlers
The robots.txt at blog.naver.com blocks GPTBot, OAI-SearchBot, PerplexityBot and ClaudeBot (measured 2026-08-03). An article published only there is invisible to ChatGPT and Perplexity, which surprises most teams entering the market.
- 03
Query language decides which documents get used
Research on multilingual retrieval reports that models systematically prefer documents in the language of the query. English-only assets tend to lose on Korean queries — which is the practical reason this site exists in Korean first.
REFERENCE
Answer engine reference
Crawler names, index sources, citation formats and how to verify each one yourself — one page per engine.
OpenAI
ChatGPT
Three separate crawlers — training, search, and user-requested
Google AI Overviews
No dedicated AI crawler — Googlebot is the whole story
Perplexity AI
Perplexity
The most visible sourcing and the fastest to reflect change
Anthropic
Claude
Separate crawlers for training, search and user requests
Naver
Naver AI Search
Heavy on domestic documents — the channel you publish on decides the outcome
Microsoft
Copilot
Runs on the Bing index — the entry point is Webmaster Tools
Gemini
Grounding is controlled by the Google-Extended token
Start with the comparison table if you want all 7 side by side.
IN KOREAN
The Korean site carries considerably more: a paper library summarising peer-reviewed and preprint research on GEO and AEO, our published measurement methodology, first-party data on AI crawler activity, and measured results from Naver search across all 17 Korean provinces. It is being translated page by page.
한국어 사이트 →Want to know how your brand appears in AI answers right now?
Send us a URL. Our free audit asks real questions in all 7 engines and reports mentions and citations separately, with the measurement conditions attached. Enquiries in English are welcome.
We reply within one business day.