The verdict in three sentences
In Tokyo, a family preparing a relocation often opens an AI assistant like ChatGPT before any school website, and asks for a shortlist based on curriculum, neighbourhood and budget. The schools that show up in those answers share consistent, dated public information (curricula actually offered, accreditations actually held, fees, transport) that third-party sources repeat. Measuring your place in those answers takes a dated record of the same questions over time, and an initial diagnostic costs roughly 3,000 to 9,000 USD on the market.
What do families ask AI assistants before choosing a school in Tokyo?
The questions no longer look like keywords. They are full sentences, often typed in London, Singapore or New York a few months before arrival. Here are realistic phrasings that keep coming back in this kind of search:
| Question put to the AI assistant | What the family really wants | Page that must answer it |
|---|---|---|
| "Which international schools in Tokyo offer the IB all the way to the Diploma?" | Curriculum continuity after a move | Curriculum page, current and precise |
| "British school near Hiroo or Azabu with a school bus?" | Daily commute for an 8-year-old | Transport page with routes and zones |
| "How much does an international school in Tokyo really cost in 2026, including entry fees?" | Total budget, not only annual tuition | Full fees page, in JPY |
| "My child doesn't speak English, which school accepts them in primary?" | Language support (EAL) | Admissions and support page |
| "Are there mid-year places for a January arrival?" | Off-cycle admissions | Admissions page with calendar |
| "Which school in Setagaya has strong learning support?" | Support for additional needs | Dedicated, factual page |
| "American school vs IB school in Tokyo for a middle schooler" | Difference in curriculum and pathways | Curriculum and approach page |
| "When are open days for international schools in Tokyo?" | Planning a visit during a look-see trip | Dated events page |
One point matters a lot: most of these questions combine curriculum, neighbourhood and logistics. A school can have a strong reputation and still be absent if its pages don't clearly say it serves Meguro, or that its programme runs through to the Diploma.
Why some schools appear in answers and others don't
AI assistants like ChatGPT build their answers from what is published, repeated and consistent. For an international school in Tokyo, five factors visibly carry weight:
- Consistent public information. If the website lists one curriculum, a parents' association directory another, and a relocation guide an old address, the answer turns cautious or names another school.
- Third-party mentions. Relocation guides, local English-language press, school association pages, accreditation bodies: they often act as the confirmation.
- Clear, factual pages. Fees in JPY, admissions calendar, bus routes, entry ages. A 40-page PDF prospectus reads poorly; a clear page is easy to cite.
- Structured data. Markup for the school, address and events helps connect the facts, without being a magic wand.
- Freshness. A fees page dated 2023 raises doubt. Families ask for this year's figures.
What not to do matters just as much: never publish exam results, rankings or accreditations the school does not actually hold, nor approximate pupil numbers. An AI assistant repeats what is written, errors included.
What this work covers, and what it costs on the market
AI visibility work is still new. In broad terms it covers an audit of what AI assistants say about the school today, a clean-up of factual pages (curricula, fees, admissions, transport), consistency across third-party listings and directories, structured markup, then dated monitoring. On data, Japan's Act on the Protection of Personal Information (APPI), overseen by the Personal Information Protection Commission (PPC), argues for minimising children's data in forms and publishing pupil photos only with parental consent. This is not legal advice.
| Service (2026 ballpark) | Range in USD | Indicative local equivalent |
|---|---|---|
| Initial diagnostic of AI answers (30 to 60 questions) | 3,000 to 9,000 | about 450,000 to 1,350,000 JPY |
| Clean-up of factual pages, English and Japanese | 6,000 to 20,000 | about 0.9 to 3 million JPY |
| Monthly monitoring of answers and adjustments | 1,500 to 5,000 per month | about 225,000 to 750,000 JPY |
| Work on third-party mentions (guides, associations) | 2,000 to 8,000 per quarter | about 300,000 to 1.2 million JPY |
| Admissions portal or student information system | on quote | varies |
These ranges reflect typical agency and consultant quotes in 2026; JPY conversions are approximate (about 150 JPY to 1 USD). At Kolonell this work is priced on quote after a diagnostic, and any Japanese and English website falls under the multilingual tier, with languages set in the quote.
How to measure without fooling yourself
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Serious measurement relies on a dated record: the same list of questions, asked at regular intervals, tracked on three indicators. The share of answers that mention the school. The accuracy of what is said (curriculum, fees, neighbourhood). And which direct competitors appear in its place. No serious provider promises that an AI assistant will recommend your school: answers vary, and the record exists precisely to show the trend rather than a single result.
Mini case study
Illustrative example: the International School of Kanda Hills, a fictional 600-place IB school in Tokyo. Across 40 family questions, the school is mentioned in 6 answers, or 15%. In 3 of those 6, the fees quoted are two years old. Leadership funds a diagnostic at 6,000 USD and an update of its fees, admissions and transport pages at 12,000 USD, 18,000 USD in total. If a single extra family enrols thanks to better presence, with annual fees of about 3 million JPY (about 20,000 USD), the spend is covered in the first year; over a five-year stay, one family is worth close to 100,000 USD. The record six months later will show whether the share of mentions rises, with no certainty in advance.
FAQ
How many questions should an international school in Tokyo track?
Between 30 and 60 questions usually cover curricula, neighbourhoods, fees and admissions. Beyond 80, monitoring gets expensive without adding much information.
Do we need to translate the whole site into Japanese?
Expatriate families mostly search in English, but Japanese returnee families search in Japanese. A Japanese version of the key pages (curricula, fees, admissions, often 8 to 12 pages) is a sound starting point.
Can we be visible in AI answers without publishing fees?
It is harder: in our table, 1 question in 8 is directly about fees, and others touch on them. Publishing a clear grid in JPY, even as ranges, helps answers stay accurate.
How long before we see a change?
Allow 3 to 6 months, as a rule, between updating pages and a measurable shift in answers. A monthly dated record lets you confirm it, or not.
Do pupil photos help visibility?
They help families visiting the site, not AI assistants directly. They should only be published with written parental consent, in line with the spirit of the APPI.
Find out how AI assistants describe your school. We record what AI assistants like ChatGPT say about your school today across some forty family questions, and we price the fixes. Request a diagnostic quote: reply within 48 hours.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.