The Three Pillars of AI Visibility: Discovery, Agentic, and Citation

Why AI Visibility Needs a New Framework

Traditional SEO gave us a single score: your Google ranking. But in the age of AI assistants, visibility is no longer one-dimensional. When a user asks ChatGPT, Perplexity, or Claude a question, the AI does not just return ten blue links. It synthesises information from across the web and decides which sources to cite, which tools to invoke, and which content to trust.

That decision process touches three fundamentally different capabilities, and most businesses are only paying attention to one of them. At Citability, we measure AI visibility across three distinct pillars: Discovery, Agentic Readiness, and Citation Readiness. Together, they produce a score from 0 to 100 that reflects how prepared your domain is for the AI-first web.

Pillar 1: Discovery (Can AI Find You?)

Discovery measures whether AI crawlers and language models can locate, access, and understand your content. Think of it as the foundation: if AI cannot find you, nothing else matters.

What Discovery evaluates:

• llms.txt presence and quality — The llms.txt file is the AI equivalent of a robots.txt. It tells language models what your site is about, what pages matter, and how to navigate your content. Sites with a well-structured llms.txt see measurably higher AI citation rates.
• robots.txt AI crawler rules — Are you allowing GPTBot, ClaudeBot, PerplexityBot, and other AI crawlers? Our scanner checks for explicit allow/block rules across 17 known AI user agents.
• Schema.org JSON-LD markup — Structured data helps AI systems extract facts about your business, products, and content. We evaluate the presence and completeness of your JSON-LD across Organization, Product, FAQ, Article, and other schema types.
• Sitemap accessibility — Can AI crawlers discover the full breadth of your content through your sitemap?

Real-world example:

A SaaS documentation site we scanned had excellent content depth and strong domain authority, yet scored just 12/50 on Discovery. The reason: their robots.txt blocked all AI crawlers by default, and they had no llms.txt. After adding four lines to robots.txt and creating a basic llms.txt, their Discovery score jumped to 38/50 within a single scan cycle.

Pillar 2: Agentic Readiness (Can AI Act on Your Behalf?)

Agentic readiness is the frontier pillar — the one most businesses have not heard of yet. As AI moves from answering questions to completing tasks, websites need to expose capabilities that AI agents can invoke programmatically.

What Agentic Readiness evaluates:

• MCP (Model Context Protocol) manifest — Does your site declare machine-callable endpoints via MCP? This allows AI agents to book appointments, check inventory, process orders, or query your API without human intervention.
• SSE (Server-Sent Events) capability — Real-time streaming endpoints enable ongoing agent-to-server communication, essential for long-running agentic workflows.
• A2A (Agent-to-Agent) protocol — The emerging standard for AI agents to discover and communicate with other AI agents. We check for agent card declarations and capability manifests.
• ai-discovery.json — An emerging protocol that consolidates AI-relevant metadata about your site into a single machine-readable file.

Why it matters now:

Agentic commerce is not hypothetical. OpenAI's Operator, Anthropic's computer use capabilities, and Google's Gemini agents are actively browsing the web and completing tasks. The businesses that expose structured capabilities through MCP and related protocols will be the ones these agents can work with. Everyone else becomes invisible to agentic workflows.

Real-world example:

An e-commerce platform we analysed scored 72/100 overall but just 4/20 on Agentic. They had strong authority, fast performance, and good content — but zero machine-callable endpoints. A competitor with lower overall authority but a full MCP manifest and SSE endpoints was already being used by AI shopping agents for price comparisons and availability checks.

Pillar 3: Citation Readiness (Will AI Cite You?)

Citation Readiness is the outcome pillar. It measures the signals that correlate with actually being cited by AI platforms when they generate responses. This is where the SE Ranking study of 129,000 domains (November 2025) fundamentally changed our understanding.

What Citation Readiness evaluates:

• Domain authority (35% weight) — The single strongest predictor of AI citations. Measured via harmonic centrality rank (hcrank), this reflects how authoritative and well-linked your domain is across the web.
• Performance (15% weight) — Core Web Vitals matter for AI too. Sites with First Contentful Paint under 0.4 seconds receive a 3x citation boost compared to slower sites.
• Content depth (15% weight) — Long-form content over 2,900 words, FAQ sections, cited statistics, and structured content all correlate with higher citation rates.
• Social proof (15% weight) — Reddit mentions, review platform presence, and community discussion signals indicate real-world authority that AI platforms weigh heavily.

The data is clear:

According to the SE Ranking study, 95% of AI citations come from organic, earned content rather than paid placements. Domain authority alone accounts for 35% of citation likelihood. A site in the top 0.1% of authority (hcrank above 8.0) is virtually guaranteed to be cited when relevant queries arise. Meanwhile, sites with excellent technical compliance but weak authority score well on Discovery and Agentic but poorly on Citation — they are ready for the future but not yet earning citations today.

How the Pillars Work Together

The three pillars are complementary, not competitive. A high-performing domain typically needs strength across all three:

What You Can Do Today

1. Run a free scan at citability.ai to see your three-pillar breakdown
2. Start with Discovery — add llms.txt and configure robots.txt for AI crawlers (the highest-ROI changes)
3. Build toward Agentic — if you have APIs or transactional capabilities, expose them via MCP
4. Invest in Citation signals — authority takes time, but content depth and performance are actionable now

Ready to see where you stand? Scan your domain free and get your three-pillar AI visibility score in under 60 seconds.