Methodology
How we measure AI visibility
The Daizie AI Business Accuracy Rubric — v3, effective 2026-07-12
Every Daizie AI Visibility Assessment produces one headline number: your AI Business Accuracy Index (0–100). This page shows you the framework it comes from and why we're the only ones publishing this much of it.
How measurement varies across the category
Most AI-visibility tools track whether AI mentions you. Daizie measures whether AI is getting you right, whether you're being recommended in the right situations, and what to fix if not.
| Most AI-visibility tools | Daizie |
|---|---|
| Track brand mentions across a few engines | Test live AI responses across four engines, then verify every factual claim against your real sources |
| Give you a visibility number with no way to change it | Every measurement has an anchored score tied to a specific fix |
| Treat every business the same | Branch by lane — a coffee shop, a coach, and a SaaS founder each get a different playbook |
| Automated dashboard, no human in the loop | Marty personally reviews every generated Assessment before it reaches you |
| Publish nothing about the methodology | This page |
| Sell you a monthly subscription for tracking | Sell you a diagnosis, an optional implementation, and optional monitoring — in that order |
Not in a vacuum
The rubric is grounded in the peer-reviewed and industry-published research on how large language models actually retrieve, rank, and cite content. Princeton's GEO paper, Anthropic's contextual retrieval work, Semrush's citation-pattern datasets, and the platform-behavior research from the AI-visibility field are baked into every driver and every band. Then — because a coffee shop and a SaaS founder don't buy the same way — the outputs are tailored to your business lane (local / services / product) so you get moves that matter for the queries your buyers actually ask.
The framework, in one paragraph
Every Assessment answers two questions in sequence. Readiness asks whether your public footprint gives AI enough evidence to describe you accurately. Measured outcomes ask what AI actually says when Daizie tests it live with the queries your buyers ask, across ChatGPT, Claude, Perplexity, and Gemini. Readiness is what you can change. Outcomes are how AI is behaving today. We measure both, separately, and never confuse them.
The five readiness drivers
Business Clarity
Can AI tell who you are, what you do, who you help, and where you work?
Source Support
Are your important claims backed by evidence AI can read?
AI Readability
Is your website structured so AI systems can actually parse it?
Distinctive Point of View
Do you have a supportable reason to be recommended over alternatives?
Recommendation Fit
Is it clear when you're the right recommendation, and when you're not?
The scoring
One headline Index. Three underlying scores.
- AI Business Accuracy Index (0–100) — the headline. Built from the three underlying scores below.
- AI Readiness Score — the five drivers, weighted, scored from public evidence. Available in the free check.
- AI Visibility Score — measured live from AI responses across the four engines. Assessment only.
- AI Business Accuracy Score — how accurately AI represents your business across six measured outcomes. Assessment only.
What's public. What isn't.
The framework, the drivers, and the research behind them are on this page. The extraction engine — the exact LLM judge prompts, the buyer-question query wordings, the fix library, the calibration data, and Marty's personal review of every generated Assessment — stays private. The framework is teachable. Applying it well is the service.
Ready to see where AI has your business right now?
- Free Readiness Check — 30 seconds, email required to see your report. Run it →
- Daizie AI Visibility Assessment — $895. Four engines, 32 responses, every factual claim verified, competitor quadrant, three readiness fixes + three accuracy fixes, 30-minute review call with Marty.
- Optimizations + Remeasure — $4,100 (add-on to the Assessment; $4,995 all-in). Marty implements or coordinates the priority fixes, re-measures across all four engines, delivers a delta report showing what moved.
- Monthly Monitoring — $197/month (add-on to the Assessment). Full 32-response protocol re-run monthly, trends dashboard, email when each report is ready.