Public data files
Use these downloads when an operator, spreadsheet, AI search visibility audit, or public report needs exact scorecard fields and thresholds.
- https://www.miragium.com/data/ai-visibility-scorecard-model.json
- https://www.miragium.com/data/ai-visibility-scorecard-model.csv
What each row contains
Each row includes metric key, label, description, input field, formula, weight, good threshold, warning threshold, weak-signal action, source review use case, source URL, and source URL.
How to use it safely
Treat the score as a directional audit, not a guarantee of AI search placement. Keep prompt sets stable, record evidence, cite public source URLs, and retest over time.
How to use this asset
Use this page as a public planning source. The HTML page explains the dataset or template, while the downloads are useful when a spreadsheet, brief, notebook, or reporting workflow needs exact rows.
- Use the page for context before copying rows into a content brief.
- Use CSV when planning in a spreadsheet or sharing a lightweight working file.
- Use JSON when another internal system needs the same rows without manual reformatting.
Limits and review notes
Public datasets are planning material. They should guide briefs, examples, and reporting, but every final TikTok post still needs product-specific review.
- Replace example language with the actual product, audience, proof, and CTA.
- Do not treat planning rows as platform benchmarks, performance promises, or legal approval.
- Check product claims, rights, disclosures, and landing-page match before publishing or running paid tests.
Public dataset
Dataset downloads and sample rows
Public scoring model for AI visibility audits, covering prompt coverage, brand mentions, source mentions, source accuracy, answer accuracy, competitor pressure, source coverage, and freshness. This page is the landing page; use the downloads below for spreadsheets, public gists, notebooks, and reporting workflows.
- Dataset ID
- miragium-ai-visibility-scorecard-model
- Rows
- 8
- Updated
- 2026-06-16
| metricKey | label | weight | goodThreshold | canonicalUrl |
|---|---|---|---|---|
| prompt-coverage | Prompt coverage | 0.1 | 20 or more prompts | https://www.miragium.com/assets/ai-visibility-scorecard-model |
| mention-rate | Brand mention rate | 0.2 | 60% or higher | https://www.miragium.com/assets/ai-visibility-scorecard-model |
| citation-rate | source mention rate | 0.2 | 45% or higher | https://www.miragium.com/assets/ai-visibility-scorecard-model |
| source-accuracy | Correct source rate | 0.15 | 70% or higher | https://www.miragium.com/assets/ai-visibility-scorecard-model |
FAQ
What does this scorecard model measure?
It measures prompt coverage, brand mention rate, source mention rate, correct source rate, answer accuracy rate, competitor pressure, source coverage, and freshness.
Does the model predict AI rankings?
No. It is an audit model for observed AI answers. It helps prioritize search and discovery actions, but AI systems decide retrieval and source mentions independently.
Can teams cite this model?
Yes. Use the HTML page for context, JSON for exact rows, CSV for spreadsheets, and source summaries for source retrieval.
Related pages
Score AI search visibility from prompt tests, brand mentions, source mentions, source accuracy, competitor pressure, and answer quality. Free, no signup required.
A public prompt suite for testing whether AI assistants mention and cite Miragium across TikTok slideshow, photo carousel, ecommerce, comparison, and operator-retrieval questions.
A public first-party source map mapping Miragium SEO pages, source summaries, topical hubs, datasets, and data downloads for humans and teams.
A public answer-block template dataset for writing concise AI Overview, answer-engine, AI search visibility, and source-backed SaaS page summaries.