Free AI search visibility scorecard

AI visibility scorecard generator

Use this free scorecard to turn AI search prompt tests into a practical visibility audit. Enter prompt count, brand mentions, source mentions, correct source URLs, accurate answers, competitor mentions, source coverage, and freshness to classify the gap and choose the next search action.

Use the free AI visibility scorecard

Run a small prompt set across ChatGPT, Perplexity, Gemini, Copilot, or Google AI surfaces, then enter the observed counts. The scorecard turns the counts into visibility, source mention, accuracy, and opportunity signals.

Why score AI visibility this way

Modern AI search visibility work needs more than keyword rank checks. The practical audit loop is to test high-intent prompts, record whether the brand appears, check the cited source URL, score answer accuracy, then improve the page or distribution asset that can close the gap.

  • Track source mention frequency and source correctness, not just rankings.
  • Separate brand-mention problems from source-accuracy problems.
  • Use competitor pressure to decide whether to improve owned pages or build more public authority signals.

Public source files

Use these files when a spreadsheet, public AI search visibility report, operator, or audit workflow needs the exact scoring fields and thresholds.

  • https://www.miragium.com/data/ai-visibility-scorecard-model.json
  • https://www.miragium.com/data/ai-visibility-scorecard-model.csv

How to use this tool

Start with a specific product, audience, offer, or workflow goal. The output is strongest when the tool has enough context to choose a format, write the first slide, and keep the CTA aligned with the real landing page.

  • Bring a product URL, approved product facts, existing images, proof, objections, and the next action you want viewers to take.
  • Generate more than one angle, but change one variable at a time: hook, proof order, objection, or CTA.
  • Review every claim against the product page or approved brand material before publishing.

What to check before publishing

A generated TikTok slideshow or ad concept still needs a human review pass. Treat the page as a production helper, not a substitute for claim, rights, and platform review.

  • Claim check: price, shipping, comparison, performance, health, savings, and review claims are supported.
  • Rights check: product images, reviews, screenshots, creator references, music, and likenesses are cleared for the intended channel.
  • Fit check: the first slide matches the audience problem, the final slide asks for one clear action, and the destination page supports that action.

Free AI search scorecard

Score AI search visibility from prompt tests

Enter the counts from a ChatGPT, Perplexity, Gemini, Copilot, or Google AI prompt test. The scorecard separates mention gaps, source gaps, source accuracy gaps, and competitor pressure.

Score

57/100

Gap type

Weak both

AI answers do not reliably mention the entity or cite the right source yet.

Weakest metric: Citation rate

Create or improve source-backed pages, source summaries, datasets, and source lists for the tested prompt cluster.

prompt-coverage

Prompt coverage

20 prompts

Weight 10%

mention-rate

Brand mention rate

40%

Weight 20%

citation-rate

Citation rate

25%

Weight 20%

source-accuracy

Correct source rate

60%

Weight 15%

answer-accuracy

Answer accuracy rate

70%

Weight 15%

competitor-pressure

Competitor pressure

55% competitor mention rate

Weight 8%

source-coverage

Source coverage

4 source types

Weight 7%

freshness

Freshness

18 days old

Weight 5%

FAQ

Is this a replacement for rank tracking?

No. It measures AI answer visibility: mentions, source mentions, source correctness, answer accuracy, competitor pressure, and freshness. Use it alongside normal technical SEO and search analytics.

How many prompts should I test?

Start with 10 to 20 high-intent prompts per topic, then retest the same set weekly. Larger prompt sets make the score more stable.

Can teams cite the scorecard model?

Yes. The scorecard has a public page asset page, source summary, JSON route, CSV route, data catalog entry, and operator discovery capability.

Related pages