Entities
Salience, disambiguation, canonical names, and linked references.
Methodology
How the AI Mention Scorecard works.
Our GEO/AEO rubric reverse-engineers how answer engines crawl, parse, and cite content. It is not a keyword checklist--it is a technical evaluation of how ingestible and verifiable your page is for modern LLMs.
Research-driven process
Six stages keep the rubric honest.
- Intent discovery: Map the high-value AI prompts your buyers ask.
- Crawl & capture: Fetch HTML, schema, feeds, and evidence exactly like AI agents do.
- Rubric scoring: Apply the 9-point GEO/AEO framework with weighted signals per engine.
- Entity & schema modeling: Align taxonomy, IDs, `sameAs`, and references.
- Answerability & evidence: Structure cite-able copy, proof, and ingestion cues.
- Implementation & governance: Provide tickets, JSON-LD, and monitoring hooks.
Rubric pillars
Every category ladders up to "Can AI engines cite you?"
Schema
JSON-LD validity, ID coverage, graph completeness, and context alignment.
Answerability
Prompt coverage, clarity, scannable headings, and structured responses.
Verifiability
Citations, evidence density, external corroboration, and trust cues.
Ingestion
Feeds, sitemaps, internal links, render fidelity, and language alternates.
Authority
Reputation signals, author entities, and cross-domain mentions.
Optimization KPIs
We measure outcomes, not vanity metrics.
- AI Intent Coverage (AIC)
- Share of tracked prompts where your brand appears in answers or source panels.
- Citation Rate (CR)
- Percentage of observed answers that explicitly cite or mention your brand/domain.
- Time-to-First-Mention (TTFM)
- Median time from publish/update to first observed AI mention.
- Entity Health Score (EHS)
- Validity + completeness of priority entities, relationships, and schema nodes.
FAQ
Methodology questions
What is the score actually measuring?
It measures how easy your page is for AI systems to retrieve, parse, trust, and cite. The score is not a generic SEO number or a traffic forecast.
Why focus on entities, schema, and evidence?
Because those are the structural signals that help models understand who the page is about, what claims it makes, and whether those claims are grounded enough to reuse in an answer.
Why is answerability a separate category?
Ranking is not the same as citability. A page can rank well and still be hard for an LLM to summarize cleanly. Answerability checks whether the content is organized for direct reuse.
Is the rubric the same for every engine?
No. The framework is consistent, but different engines emphasize different retrieval, formatting, and grounding behaviors. That is why the rubric uses weighted signals rather than a one-factor checklist.
Does the score replace human review?
No. The score is a decision tool. It helps prioritize where to edit, but strong editorial judgment and subject-matter accuracy still matter.
Want the deep dive?
We publish experiments, release notes, and templates.
The Lab section (coming soon) will house KPI experiments, prompts, and entity templates so your team can keep pace with model updates.