# aitani — AI visibility for Japanese local businesses (expanded reference) > aitani is a managed service that helps local businesses in Japan get > discovered by AI assistants and local search: ChatGPT, Gemini, Google AI > Overview, and Google Maps. This llms-full.txt expands on llms.txt with the > full measurement method, funnel definitions, adapter list, and operating > constraints. Independent service, not affiliated with Google or OpenAI. ## About aitani - Founded 2025 by Khoa Nguyen (founder; ex-Google, ex-SoftBank; Master's, University of Tokyo). - Trilingual service: English, Japanese, Vietnamese. - Areas served: Japan (nationwide, prioritising Tokyo, Chiba, Aichi, Kanto metro). - Business model: free AI Search Check audit; a managed monthly plan and an optional Launch Setup add-on for new businesses, both quoted per business after a scoping conversation. No list price is published; the contractual price is disclosed before signing and is stated in the Terms and in the Act on Specified Commercial Transactions notice. ## Core proposition aitani measures how ChatGPT, Gemini, Google AI Overview and Google Maps present a specific local business, then — for paying customers — implements improvements across sources the business controls (website, Google Business Profile, listings, reviews replies) and re-measures the change on the same question set. ## Measurement method ### The four-tier funnel aitani evaluates visibility as a funnel with four levels, from broadest to narrowest: 1. **Knowledge** — Does the AI know the business exists at all? Detected by asking name-anchored questions ("What can you tell me about X in Y?") and parsing whether the AI returns real attributes vs. hallucinated or refused. 2. **Category** — Is the business associated with the categories real customers search for (e.g. "family-friendly ramen", "halal restaurant", "open past 22:00", "English menu")? Detected by asking category-anchored questions and parsing whether the business is enumerated. 3. **Consideration** — When a customer asks an open-ended local question, does the AI place this business inside the shortlist it presents? Detected by asking decision-oriented questions and parsing shortlists. 4. **Recommendation** — Is the business picked as the top or explicit suggestion? Detected by parsing rank / emphasis / "I recommend" language. Gains at Recommendation without gains at Knowledge and Category are fragile and typically regress within 30–60 days. The whole funnel must move. ### Question set For each business we compose 45–50 customer questions covering: - Intent: browse, decide, book. - Audience: families, tourists, groups, solo, students, business. - Constraints: allergy, budget, late-night, accessibility, language. - Language: Japanese primary, English secondary, Vietnamese where relevant. The question set is locked at baseline. Every re-measurement uses the same set. Comparing different sets is disallowed — it produces meaningless deltas. ### Query execution - **ChatGPT** — OpenAI Responses API with web browsing enabled, from a Japanese IP address. Five repeats per question. We record the modal answer plus variance. - **Google AI Overview** — DataForSEO SERP endpoint with location_code and language_code forced to Japan / Japanese. Five repeats per question. - **Google Maps rank tracking** — DataForSEO Maps rank tracking or Playwright from a Japanese IP for spot-checks. - **Google Places / Business Profile** — Google Places API for public attributes; Google Business Profile API (via OAuth from the owner) for private metrics like impressions, calls, direction requests. - **PageSpeed / Core Web Vitals** — Google PageSpeed Insights API. - **Gemini and Google AI Mode** — currently not machine-measurable at scale. We do not fabricate a score for these; we record "not measured" and rely on human spot-checks documented in the workspace. ### What counts as "recommended" A mention ("Sakura Ramen is one of many ramen shops in Shibuya") is not the same as a recommendation. We only credit a hit when the business appears in an enumerated shortlist the AI presents in response to a decision-oriented question. This is stricter than most AI-visibility tools on the market. ### White-hat principle We only clean up and align real evidence about a business: - Consistent NAP (name, address, phone) across every source. - Consistent opening hours across GBP, website, third-party listings. - Structured schema (LocalBusiness, Restaurant, FAQPage, Menu). - Plain-text menus (not menu-in-a-photo). - Factual, non-embellished business descriptions. - Honest photo captions. - Reviews reply strategy that acknowledges real feedback. We do not: - Create fake reviews or purchase reviews. - Buy backlinks or use link farms. - Inject hidden text or cloaking. - Publish claims we cannot substantiate. Where evidence cannot be obtained, we mark the item "not measured" rather than guess. Not-measured items never inflate the score. ## Operating limits - **Per-run spend cap** — default USD 3.00 per audit. Enforced at the adapter level; a run stops rather than exceed the cap. - **Per-day account cap** — global daily cap enforced across all runs so no runaway API bill is possible. - **Rate limiting** — per-IP request limiting on the intake endpoint; Turnstile challenge on suspected abuse. - **PII scrubbing** — logs and error traces are scrubbed of email, phone and free-text notes before being sent to Sentry. ## Adapter list Data providers currently integrated: - OpenAI Responses API — ChatGPT answers. - DataForSEO — Google AI Overview, Google SERP, Maps rank, Business Listings. - Google Places API — public attributes. - Google Business Profile API — private metrics (OAuth from owner). - Google PageSpeed Insights API — Core Web Vitals. - Resend — outbound email. - Turnstile — bot protection on intake. Data processors (infra): - Vercel — hosting, edge functions (Tokyo hnd1). - Neon — PostgreSQL (Singapore ap-southeast-1). - Inngest — background job orchestration. - Sentry — error monitoring. ## Product - [Homepage (English)](https://www.aitani.io/en) - [Homepage (Japanese)](https://www.aitani.io/ja) - [Homepage (Vietnamese)](https://www.aitani.io/vi) - [Pricing](https://www.aitani.io/en/pricing) — managed plan, quoted per business - [About](https://www.aitani.io/en/about) — founder profile - [Case studies](https://www.aitani.io/en/case-studies) — in preparation. When published, every figure on these pages states the platform, the date and the method it was measured with, and any relationship between aitani and the business is disclosed on the page. - Industry pages, one per industry aitani serves — restaurants & cafés, hotels & travel, clinics, salons, professional services. Each states what usually goes wrong before a customer arrives, the conditions customers actually compare on, the six areas aitani audits, station-level and catchment-level visibility, and how the work runs. https://www.aitani.io/en/services/{restaurant,hotel-travel,clinic,salon,professional} The restaurant page carries measured figures from a real engagement (a yakiniku restaurant near Sannomiya, Kobe, measured 2026-08-13, method 2.7-geo-real); the business is not named because consent to name it has not been given. The other four industry pages carry sample figures and say so. - Product pages, one per product. https://www.aitani.io/en/products/{free-check,aio-geo-llmo,seo,meo,integrated} All figures on the product pages are sample data, labelled as such. - [Methodology](https://www.aitani.io/en/methodology) — human-readable method summary - [Blog](https://www.aitani.io/en/blog) — coming soon ## Legal - [Terms of Service](https://www.aitani.io/en/terms) - [Privacy Policy](https://www.aitani.io/en/privacy) - [特定商取引法に基づく表記 (Act on Specified Commercial Transactions)](https://www.aitani.io/ja/tokushoho) ## Reproducibility Every audit stores: the exact question set (versioned), method version, model version, IP region, timestamps, raw responses, and the derived score. If a customer disputes a result, we can replay the exact same query and show them the response. This is why we invest in append-only logging. ## Contact - General: aitanijapan@gmail.com - Privacy / deletion requests: aitanijapan@gmail.com - Media / partnerships: aitanijapan@gmail.com ## Reference - [llms.txt](https://www.aitani.io/llms.txt) — short version - [Sitemap](https://www.aitani.io/sitemap.xml)