Rating Methodology
How Korea AI scores product impact, global competitiveness, defensibility, breakout efficiency, AI centrality, and evidence quality.
Five independent signals
Impact
Product-specific adoption, workflow depth, and reach. Domain importance and parent-platform exposure do not count.
Global
International adoption, named overseas customers, regulatory reach, and infrastructure or developer ecosystem fit.
Protected
Workflow depth, proprietary data, switching costs, regulated deployment, hardware constraints, and durable networks.
Breakout
Product quality and adoption efficiency normalized for company size, so emerging teams can surface without pretending to be larger.
AI Core
How central AI is to the product rather than a minor feature added to a larger non-AI surface.
Evidence comes first
A score is only as useful as the public evidence behind it. Evidence badges qualify confidence; they never secretly change the number.
Sourced
Official or credible public sources plus adoption evidence or a written score rationale.
Partly sourced
At least one source, adoption signal, or rationale, but not enough to independently support the full assessment.
Unverified
No source, product-specific adoption signal, or individual rationale. Treat the score as a catalog placeholder.
Non-negotiable caps
Adoption
Parent-platform reach, bundled exposure, downloads, API calls, and unnamed enterprise labels do not become product users.
Regulated niches
Clearance is a defensibility signal, not Impact. With no equivalent deployment proof, userCount 0 products are capped at 78.
AI hardware
No named deployment caps Impact at 78; one regional deployment at 82; multi-partner or quantified volume at 86; global standard is uncapped.