How does your business come across
in your customers' language?
When a traveler asks "where can I do a tea ceremony in Kyoto?", the answer increasingly comes from AI rather than Google — usually as a list of three to five places. What separates the businesses that get chosen from the ones that don't? I test this myself — in Japanese, English, Chinese, and more — and analyze the results.
From the July 2026 survey
I asked AI about ten cultural-experience businesses in Kyoto, in Japanese, English, Traditional and Simplified Chinese. A shop with over 500 positive reviews on Google Maps was said to "not exist", and was then described as a different business entirely — while the one business whose prices AI quoted accurately was a small, unassuming tea room.
Where does that difference come from?
About this lab
Kääselä (Kamogawa AI Search Lab) is an independent research project that measures how generative AI recognizes and describes businesses, and publishes the records.
This field goes by several names — GEO (Generative Engine Optimization), LLMO, AIO — and is new enough that none of them has settled. Chasing the latest name — or adding another word with no substance behind it — is not the point. The question I care about is what it takes for a business to be properly recognized by its customers in the age of AI — and the goal of this lab is to answer that with data.
It is run by one person: an IT consultant living in Kyoto, working on this alongside a day job. Not affiliated with any university or organization.
Contact
kamogawa.ai.lab@gmail.com
Currently employed full-time; not accepting paid engagements. Happy to share findings and methodology at no cost, where it aligns with the purpose of this research.