What GEO Is · How We Score It

Stop being searched.
Start being recommended.

First, measure — for free. Understanding can come later.

GEO (Generative Engine Optimization — LLMO in Japan) is the discipline that gets generative AI to read your business correctly, so it can recommend you. And the ruler that measures it, 0 to 100.

How AI Reads

Server access log for je-jo.com showing GPTBot, ClaudeBot and PerplexityBot requests over 30 days
01The crawler reads.

GPTBot, ClaudeBot, PerplexityBot crawl the site. Getting into AI at all is the starting point.

A plate of je-jo pasta beside that restaurant's Restaurant JSON-LD — hours, address, cuisine, rating
02Structured data carries meaning.

Hours, menu, location — encoded as JSON-LD and llms.txt, machine-readable. AI can cite them without guessing.

Three AI answer cards in Japanese, English and Chinese naming JE&JO, over the measured discovery rate per engine
03In every language, an answer.

Travelers ask AI in their own language. Prepared in Japanese, English, and Chinese, you become the answer to more of those questions.

City lights spreading at dusk — reaching across languages

Multilingual × AI Visibility

An Honest Distinction

We measure the two separately — because a company that conflates them can dress up what it built as a result it earned.

Readiness
Is it readable to AI?

Structured data. Crawler access. Multiple languages. We can build this ourselves — which is why we can promise it.

Realized
Did AI actually name you?

The result of a live test across 3 engines × 3 languages. Not something you can build — only something that shows up — so we measure it honestly, every month.

Our Score · Diagnostic Metrics

The GEO score is one deterministic ruler, 0 to 100. We publish the methodology, the weighting, even our own score. Would you trust a company whose builder hides its own number?

Measure your own score, free

Scoring Criteria

01

Can AI crawlers get into the site (robots.txt / llms.txt)

02

Is business info and the menu machine-readable via JSON-LD

03

Are Japanese, English, and Chinese — and hreflang — correctly set up

04

Is the content specific and fresh enough that AI wants to cite it

05

Were you actually named — tested across 3 engines × 3 languages

Detailed scoring criteria and pass/fail examples for each station will be published on a rolling basis.

Check for free
Free AI check