Skip to content

CASES

What we built,
and what we measured

We have no client case to show, so we made our own work the subject. Statuses below are written as they are — measurement in progress is not dressed up as a result.

⚠️ DECLARED UP FRONT

Every case on this page is Navirang's own work — a service we built, our own site, and this site. None of them is a client engagement, and none is a controlled experiment. The sample is one company, and there is no guarantee that what we observed reproduces elsewhere. We publish the measurement conditions and the raw data so you can check and judge for yourself.

PRODUCT BUILD

Built · re-measurement planned

Reviewloger — from product design to AI search diagnosis

A project that started from a problem: brand-ambassador campaign listings are scattered across platforms and cannot be found in one place. Navirang designed and built the service, and the collection pipeline, data normalisation, search structure and SEO/AEO structure were all made by one team.

Collection adapters
62 implemented (61 currently active)
Listings searchable
6,147 as of 2026-08-12
Scope
Service design · web build · data collection · content structure · SEO · AEO · structured data · measurement
Measurement status
AI search discoverability diagnosis complete. Improvement results will be published separately after re-measuring the same question group

WHAT THIS SHOWS

  • That we can design and build a service, not only run marketing for one
  • That backend work such as collection and normalisation sits inside the same design as AEO
  • That a good service is not automatically discovered by AI — confirmed by measurement
reviewloger.com ↗

OUR OWN SITE

Baseline measurement in progress

Our own eight-week log — 40 questions × 7 answer engines

We made our own site the subject, because we have no client case to show. A company selling AEO measures its own citation status every week and publishes the record as it is, including the weeks the numbers get worse.

Unit of measurement
40 questions × 7 answer engines = 280 cells
Conditions
Same day and time of day each week, logged-out sessions, question wording fixed
Published in full
The question set · the measurement protocol · how the citation rate is aggregated, denominator included · the raw data
Current status
Baseline measurement of 280 cells in progress. No round is published before the baseline is complete

WHAT THIS SHOWS

  • That measurement conditions are published before there are results
  • That we do not write 'improved' without a starting point
  • That the conflict of interest — measuring ourselves — is declared before we begin
The measurement protocol →

CLINIC BUILD

Built · awaiting deployment

jeongseokps.com — making a plastic surgery clinic legible to answer engines

A new site for Jeongseok Plastic Surgery in Gangnam, Seoul. On the previous site the body copy lived inside flattened images, so neither a search engine nor an answer engine could read a single character of it. Navirang rebuilt the site and designed the human-facing pages and the machine-readable layer together.

Scope
Site build · content structure · SEO · AEO · GEO · structured data · discovery infrastructure
Structured data
Clinic (MedicalClinic) · site (WebSite) · pages (MedicalWebPage) · surgeons (Physician) · concerns (MedicalCondition) · procedures (MedicalProcedure) · FAQ (FAQPage) · trail (BreadcrumbList)
Discovery infrastructure
Sitemap with per-page lastmod, one source for every canonical, a documented robots policy, share cards, and a generated llms.txt
Measurement
Not yet deployed. There are no results to report; measurement begins after launch

WHAT THIS SHOWS

  • That AEO is a question of site structure rather than of writing
  • That in a heavily regulated field, what a site declines to say is part of the design
  • That the machine-readable layer is built alongside the site, not bolted on afterwards

THIS SITE

Live

navirang-ai.com — applying the AEO standard to our own site first

If a site selling AEO is itself blocked, the argument does not hold. This site is the result of applying the same standards we recommend to clients, and most of what you see is a working example of them.

Rendering
Static output (SSG) — body text is in the HTML source
The machine-readable layer
Per-page JSON-LD, a sitemap with per-page lastmod, robots.txt allowing AI crawlers explicitly, llms.txt (optional)
Content structure
Question-shaped headings · a direct answer in the first paragraph · FAQs that stay in the source when collapsed · stated update dates
Published data
340 measured items across all 17 Korean provinces, with the raw CSV at a public address

WHAT THIS SHOWS

  • That building a website and doing AEO are not separated here
  • That we hold to the standard we recommend, first
  • That we produce and publish primary data ourselves
Answer engine reference →

The full build logs and the week-by-week series exist in Korean only for now. The Korean site carries the complete records, the raw CSV addresses, and the measurement series. Cases (Korean) →

Want the same measurement run on your site?

Send us a URL. The free audit records the current state with its conditions attached, so there is a starting point before anything is called an improvement.

Request a free AEO audit 070-4015-6848

We reply within one business day.

Free audit Call Email Blog