Key takeaway: trensee AIVS measured six Korean industry-leading brand pages and found an average score of 60 (C grade). Only one brand reached B grade. FAQPage schema and llms.txt adoption were both 0%. As AI search becomes mainstream, Korean brands' AI visibility readiness is still at an early stage.

Why does AI visibility matter right now?

Usage of AI search tools — ChatGPT, Perplexity, Claude, Gemini — is growing rapidly. When a user asks an AI "What is the transfer fee for this fintech app?", whether that brand's official page is cited as the answer source is determined by entirely different criteria than search engine rankings.

This benchmark used trensee's AI Visibility Score framework (AIVS) to directly measure the service guide and FAQ pages of representative Korean brands across six industries. Brand names are anonymized to prevent claims. Measurement date: March 18, 2026.

How was AI visibility measured?

How is the score calculated?

trensee AIVS is built on four pillars:

Pillar	Max Score	What Is Measured
Authority	59 pts	External citation quality, FAQ count, content length, FAQPage schema
Readability	20–28 pts	Title/meta description optimization, question-heading ratio
Structure	46 pts	BreadcrumbList schema, heading count, llms.txt, image alt coverage
AI Infra	15 pts	AI crawler access (GPTBot, ClaudeBot, PerplexityBot), page speed

The four-pillar total is normalized to 100 points and graded A (90+) · B (75–89) · C (60–74) · D (45–59) · F (44 or below).

Why were service guide pages measured instead of homepages?

Homepages (app download / marketing landing pages) contain minimal content by design, producing structurally low scores. This benchmark targeted service introduction pages, usage guides, and help center FAQ pages — the content AI actually uses as answer sources.

What scores did the six industries receive?

What were the overall results?

Industry	Grade	AIVS	Authority	Readability	Structure	AI Infra
Fintech unicorn	B	76	54%	89%	72%	93%
Large messenger / platform	C	66	49%	85%	52%	47%
E-commerce platform	C	62	44%	82%	48%	53%
Delivery / O2O platform	D	54	36%	75%	37%	40%
Used goods marketplace	D	51	32%	71%	40%	47%
Gaming / global content	D	48	27%	68%	35%	53%
Average	C	60	40%	78%	47%	56%

How were grades distributed?

B or above (AI citation likely): 1 brand (17%)
C grade (improvement needed): 2 brands (33%)
D grade (foundational gaps): 3 brands (50%)
F grade: 0 brands

What differed across industries?

Fintech unicorn — B grade (76): Korea's only standout

The only brand to reach B grade. A modern tech stack (React/Next.js) and well-structured guide pages were the key strengths. All three AI crawlers were permitted, and Core Web Vitals scores were excellent.

Strengths: Fast speed, structured guide content, full AI crawler access Gaps: No FAQPage schema, no llms.txt — approximately 14 points short of A grade

Large messenger / platform — C grade (66): rich content, low AI optimization

Content volume and external media citation count were the highest among the six. However, AI crawlers were partially blocked, and FAQPage schema and llms.txt were absent. A clear case of sufficient content assets but no investment in AI visibility optimization.

Strengths: Rich content, high external authority Gaps: Partial AI crawler blocking, zero AEO structured data

E-commerce platform — C grade (62): product schema present, AEO absent

Product JSON-LD schema was well implemented on product pages. However, this schema serves shopping search — not AI answer citation. FAQPage and HowTo schemas were absent.

Strengths: Product schema, fast CDN Gaps: No AEO structure on service guide pages, no question-form headings

Delivery / O2O platform — D grade (54): app-first strategy leaves web exposed

The app is the primary channel, so web page optimization investment was relatively low. Minimal content, almost no structured data, and zero question-form headings.

Strengths: Brand awareness (external mentions) Gaps: Thin web content, no structured data

Used goods marketplace — D grade (51): modern stack, zero AEO

Tech stack was modern and page speed was adequate. However, service guide page content was short, with no FAQ structure or schema.

Strengths: Modern tech stack, speed Gaps: Short service descriptions, no FAQ, no llms.txt

Gaming / global content — D grade (48): global recognition offsets weak AEO

Brand mentions on global platforms prevented Authority from hitting zero. However, Korean-language service page AEO optimization ranked lowest among the six.

Strengths: International media mentions (global recognition) Gaps: Lowest Korean AEO score, no question-driven content

What were the common weaknesses across all six brands?

1. FAQPage schema adoption rate: 0%

None of the six brands applied FAQPage JSON-LD schema. Adding this single element can raise AIVS by up to +13 points (Authority 9 pts + Structure 10 pts, after normalization).

2. llms.txt file adoption rate: 0%

No brand provided a llms.txt file to help AI systems understand the site's purpose and key content. Low implementation cost, but delivers up to +7 points across AI Infra and Structure.

3. AI crawler access rate: 33% (2 of 6)

Four of six brands blocked at least one of GPTBot, ClaudeBot, or PerplexityBot in robots.txt. Blocking AI crawlers directly removes the site from that AI's indexing and training pipeline, lowering brand mention probability.

How can AI visibility be improved?

Step 1: Immediate (1–2 days, low dev cost)

Update robots.txt: Allow GPTBot, ClaudeBot, PerplexityBot, Google-Extended
Create llms.txt: Add service description, key URLs, and methodology at site root

Step 2: Short-term (1–2 weeks, moderate effort)

Add FAQPage schema: Apply JSON-LD FAQPage to key service guide pages
Convert headings to question form: "How to use" → "How do I get started?", "Fee guide" → "What are the fees?"

Step 3: Medium-term (1–3 months, content investment)

Expand guide content: Publish 300-word+ guide pages that directly answer user questions
Build external mentions: Tech blog posts, media contributions, and English documentation to strengthen AI training data signals

How long does it take to reach A grade by industry?

Industry	Current Score	Gap to A	Estimated Timeline
Fintech unicorn	76	+14 pts	2–4 weeks
Large messenger / platform	66	+24 pts	1–2 months
E-commerce platform	62	+28 pts	1–2 months
Delivery / O2O	54	+36 pts	2–3 months
Used goods marketplace	51	+39 pts	2–4 months
Gaming / global	48	+42 pts	3–6 months

FAQ

Q. Which brands were measured in this benchmark?

Brand names were anonymized to ensure neutral data and avoid claims. One representative brand per industry was selected — fintech, platform, e-commerce, delivery, used goods marketplace, and gaming — based on monthly active user rankings in Korea.

Q. Why were service guide pages measured instead of homepages?

Homepages are intentionally minimal in content, producing structurally low AIVS scores. AI answer engines actually cite service guides, usage documentation, and FAQ pages — so measuring these gives a more meaningful comparison.

Q. Does a low AIVS score mean the brand never appears in AI answers?

Not necessarily. AIVS measures technical optimization readiness; actual AI mentions also depend on brand reputation, AI training data inclusion, and query context. However, low AIVS increases the risk of outdated or missing brand descriptions in AI-generated responses.

Q. Can I check my own brand's AIVS score?

Enter any URL into trensee's AI Visibility Diagnostic to get a free AIVS score, grade, and per-signal breakdown. → AI Visibility Diagnostic (AIVS)

Q. How often will this benchmark be updated?

AI crawler policies and LLM model updates can shift optimization criteria, so a quarterly update cadence is the target. Next benchmark: June 2026.

Is now the right time to optimize for AI visibility?

Korean brands' average AI visibility readiness sits at C grade (60 points) — still early-stage. This is also an opportunity: brands that optimize now can establish a position before the competition solidifies.

AI search citations, like search rankings before them, are easier to win early. Three foundational steps — adding FAQPage schema, creating llms.txt, and allowing AI crawlers — are enough to surpass today's top-ranked brands in AI visibility while the gap is still closeable.

→ AI Visibility Diagnostic (AIVS)

What are the measurement conditions and limitations?

Measurement date: March 18, 2026 (subsequent brand updates not reflected)
Pages measured: One representative service introduction, usage guide, or help center FAQ page per brand
Measurement tool: trensee AI Visibility Diagnostic (AIVS, API v2)
Limitation: Live LLM brand mention testing was not included in this benchmark. It will be added in the next edition.

structured-data Posts

Korean Brand AI Visibility Benchmark — March 2026 AIVS Report