한국Global日本中国GCC무료 점수카드 →
What's happening

A buyer just asked ChatGPT for the best brand.
It named three. You were not one.

They finished the purchase without ever knowing you existed. This is not a low ranking — it is being left off the shortlist entirely.

The 5 things we measure

The checks the real report runs. Nothing is auto-scored here — the score is run by a person against your live site and emailed to you. Read what each check means first.

  • Crawler access — can the bots ChatGPT & Claude send to read your site (GPTBot, ClaudeBot) actually get in? 3 of 10 blocked the bots
  • Citable paragraphs — do you have self-contained 140–160-word blocks an AI can quote as-is? 0 of 10 had any
  • llms.txt — a plain file that tells AI how to read you (the cheapest AEO asset to add)
  • Structured data — is your company & FAQ info marked up so machines can read it (JSON-LD)?
  • Trust signals — do outside sites link to and describe you?
Why now

AI answers are winner-take-few. Links were winner-take-many.

When search returned ten blue links, being on page one was enough. AI answers name two or three brands and stop. Getting into that set is a different problem: the engine must be able to crawl your page, lift a self-contained answer from it, and trust you enough to cite it. Research on multilingual retrieval even finds a bias toward English documents regardless of relevance — so non-English brands are structurally disadvantaged unless their English surface is built for this. We measured ten export brands in one category: three were silently blocking AI crawlers at the CDN, and not one had a single answer-shaped paragraph.

Answer engines
ChatGPT · Perplexity
Search AI
Google AI Overviews
Crawlers
GPTBot · ClaudeBot
Index
Bing · Perplexity

We probe each engine's crawler with its real user-agent, and score per engine — weights differ (AutoGEO, GEO-16).

Three ways to get diagnosed

New to AEO? That's fine. There are three ways to find out whether your brand makes it into AI answers — ask the AI yourself, hire an agency, or have us measure it. It comes down to what each one gives you and what it can't.

DIY promptingAEO/GEO specialist retainerAEO Korea diagnostic
How it measuresAsk ChatGPT a few times — answer changes each runVaries by firm; method often undisclosedReproducible 5-dimension audit + 4 real crawler fetches
Detects crawler blocksUsually notDepends on the firmYes — browser vs GPTBot response compared every time
Basis for fixesPersonal experience / intuitionEach firm's in-house methodCited research (GEO, KDD 2024)
BillingFree (but not reproducible)Monthly retainer (ongoing)One-off report from $390 · flat fee
TurnaroundInstant but unstableOften weeks to onboardScorecard 2 days · report 5 business days

The comparison is to a boutique firm that specializes in AEO/GEO — not a full-service marketing agency. A monthly retainer is an ongoing engagement; our diagnostic is a one-off report, so the billing models differ. Retainer ranges are an estimate from public quotes and typical market rates.

What changes

The goal: thirty days later, the answer says your name.

Picture the re-measurement we aim for: last month 'below floor', this month 'citable'. Getting there takes accurate measurement and a few fixes — not a rebrand. We don't guarantee the AI will name you (a private algorithm makes that call); what we guarantee is reproducible measurement and a 30-day re-measure that shows the change in numbers.

Before

Now (below floor)

Crawlers blocked, zero citable paragraphs, no structured data. AI describes you from third parties only.

After · 30d

Re-measured (citable)

Crawler access restored, answer blocks per buyer question, clean entity graph. AI quotes you.

Proof

측정 → 처방 → 재측정

Every score is reproducible; every fix is annotated with the research behind it (GEO, KDD 2024). We re-measure in 30 days.

Deliverable preview

The real report, real structure

These three pages are lifted from an actual diagnostic (brand anonymized) — the same case (TC-KB-001) shown on the sample report page, with the same five dimensions and the same score, so the two never disagree. 85 is the citable floor and our measured cohort averages in the 30s, so this case's 28.6 is 'average but below the floor'. The full deliverable is a 15–25 page PDF, and every number is reconciled against a snapshot before it ships.
Page 1 · Overall score

28.6 / 100

Case TC-KB-001 · K-beauty D2C (anonymized) · citable floor 85 · measured cohort averages in the 30s → this case: average but below floor
Citability 25%0
Structure 20%30
Multimodal 15%51
Trust signals 20%20
Technical access 20%55
Page 2 · Per-engine · crawler fetch

Where you stand, engine by engine

chatgpt31.2 · below floor
perplexity28.6 · below floor
google-aio26.4 · below floor
crawler fetch (browser / gptbot / claudebot):
browser200
gptbot403
claudebot403
Page 3 · Fix roadmap

Prioritized, evidence-tagged

  1. Unblock GPTBot·ClaudeBot at CDN — instant, config
  2. Add 8× 140–160-word answer blocks — citation sweet-spot
  3. Ship llms.txt + Organization JSON-LD — hygiene
  4. 3 comparison articles — 2.4× mention lift, Semrush
  5. Third-party mentions (YouTube/Reddit) — r=0.737

See the full sample report →

Pricing

Flat fee, published

What to expect — free scorecard in 2 business days. Paid diagnostic runs request → agreement & payment link within 1 business day → you pay → report within 5 business days. It is not an instant checkout.

Prices are exactly as listed above — flat fee, no negotiation. Start with the free scorecard. For the paid diagnostic or monitoring, click the button below and we email you the agreement plus a bank-transfer payment link within one business day — you pay and we start. Payment is currently by bank transfer (service contract, 3.3% withholding); invoice and card are being added once business registration is complete. Nothing is charged until you transfer.

Free scorecard

$0
optional 30-min call
  • 5-dimension readiness score
  • Position vs your cohort
  • Single most urgent issue
Request
Less than one month of a retainer

Diagnostic report

$390
one-off flat fee · 5 business days · (intro price, $690 later)
  • Full diagnostic + bottleneck
  • Prioritized fix roadmap
  • 30-day re-measurement
  • Full refund if unmeasurable
Enquire

Monthly monitoring

$390/mo
cancel anytime
  • Monthly delta report
  • Competitor movement
  • New-issue alerts
Ask
FAQ

Questions you probably have

How is AEO different from SEO?+

SEO is about ranking in a list of links. AEO is about being cited and recommended inside the answer an AI writes. The old page of ten links is now one answer that names two or three brands, so getting into that set is a different problem.

Do you guarantee visibility?+

No. AI answer algorithms are private, so promising visibility would be a lie. We promise reproducible measurement, fixes backed by published research, and a re-measurement 30 days later with the same method. If a site can't be measured, we refund in full.

Why is this price justified?+

If one report catches a single misfiring crawler block, it undoes a state where the brand was entirely absent from ChatGPT's answers. In our cohort, 3 of 10 were in exactly that state. Unlike an AEO/GEO specialist's monthly retainer (an ongoing commitment), this is a one-off flat-fee diagnostic at $390, so it's a low-risk way to start.

Who is this for?+

Export-focused brands whose overseas buyers ask ChatGPT, Perplexity or Google's AI. K-beauty, cross-border D2C, and B2B manufacturers. If real revenue rides on being named by an AI, this is for you.

How long does it take?+

The free scorecard takes two business days; the paid diagnostic five. Some fixes, like unblocking crawlers, are a few config lines and land within days.

Who builds it? Do you actually know AI search?+

In plain terms — we built a tool that opens your site through the eyes of the programs AI uses to read the web (crawlers), and mechanically scores whether your text is shaped the way AI likes to quote. We calibrated it against a US firm in the same field (Answerability) and we don't promise what we can't measure.

We built the audit engine ourselves — code that fetches each site with the real user-agents of four crawlers (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot) and compares response codes, analyzes the length distribution of server-rendered paragraphs to count citable blocks, and parses robots.txt, llms.txt and structured data. The methodology mechanizes Answerability's (answerability.ai) audit framework, and we calibrated our engine's scores against their manual audit of the same sites so the two don't diverge. It's run by people with a data/engineering background plus a corporate-finance discipline of never promising what can't be measured — which is why you won't see us guarantee AI visibility. Business and contact details are on our Terms and Privacy pages.

Start here

Send a URL. Get a score in two days.

Leave a brand and a site URL. We return a 5-dimension scorecard and your position against the cohort, free.

Used only to send your scorecard. No drip sequences.