What is an Agentic Score?
An Agentic Score is a deterministic 0–100 rating of how ready a website is to be discovered, understood, and transacted with by AI agents and LLM-powered assistants — ChatGPT, Claude, Perplexity, Gemini, and the wave of buying agents that follows them.
The three dimensions
- Discoverability — can crawlers from OpenAI, Anthropic, Perplexity, and Google reach your pages, and do you publish an
llms.txtmanifest and sitemap? - Understandability — is your pricing, product, and company info in Schema.org JSON-LD? Is the HTML server-rendered so an agent sees content without executing JavaScript?
- Buyability — can a user (or an agent) sign up self-serve, hit a public API, and read terms/privacy without a sales call?
Why deterministic
Two scans of the same site produce the same score. There is no black-box AI grading your brand. Every point is earned or lost from a public artifact — a header, a file, a schema block — that you control. That is what makes the Agentic Score useful as a KPI: you can ship a fix on Monday and re-score on Tuesday.
How it’s calculated
Twelve weighted checks summing to 100. Full point breakdown, fetch behavior, and pass / partial / fail rules are published in the Methodology v1.0.
Agentic Score vs AI Visibility
They measure different things. The Agentic Score is a site-side measurement of what an agent can do once it arrives. AI Visibility is a market-side measurement of whether LLMs mention your brand in answers to category questions. Visibility varies weekly and by provider; the Agentic Score is stable. Fix the Agentic Score first — it’s the ground you stand on.
FAQ
What is an Agentic Score?
An Agentic Score (0–100) is a deterministic measurement of how ready a website is to be discovered, understood, and transacted with by AI agents and LLM-powered assistants (ChatGPT, Claude, Perplexity, Gemini). It is computed from 12 public, machine-verifiable checks across three dimensions: discoverability (can agents crawl and index you), understandability (can they parse pricing, products, and policies), and buyability (can they complete a purchase or hand off to a human).
Why does the Agentic Score matter for SaaS?
AI assistants are becoming the top of funnel. When a user asks 'best invoicing tool for freelancers', the assistant answers with one or two names — not a list of 10 blue links. If your site isn't machine-readable, you are invisible in that conversation regardless of your SEO ranking.
How is it calculated?
The score sums weighted points from 12 checks: llms.txt (10), AI crawlers in robots.txt (10), Schema.org JSON-LD (10), parseable pricing (12), public API docs (10), agent manifest (8), sitemap (6), meta/OG completeness (6), self-serve signup (10), machine-readable company info (6), server-rendered HTML (8), and legal pages (4). Every check is deterministic — two scans of the same site return the same score.
Is the Agentic Score the same as AI visibility?
No. The Agentic Score measures transactability — what an agent can do on your site once it lands. AI visibility measures whether LLMs cite you when asked category questions. Both matter, but the Agentic Score is deterministic and improvable this week; visibility varies by query and provider.
Who created the Agentic Score?
AIScoring (aiscoring.io) publishes the Agentic Score methodology under a public, versioned spec. The current version is v1.0.
Ready to see your score?
Run a free scan →