mentd.
mentd. methodology

AI Discovery Score

The AI Discovery Score is a 0–100 composite built from deterministic observation data. It answers one question quickly: how discoverable is this domain across modern AI systems right now?

Formula

Score = weighted average of six components. Each component is normalized to 0–100, then multiplied by its weight. Recalculates after prompt runs, audit runs, integration syncs, and scheduled recalc jobs.

Citation Rate30%

Percent of tracked answers where the domain is directly cited.

Source: visibility_observations

Mention Frequency20%

How often the brand is mentioned across the tracked AI answer set.

Source: visibility_observations

AI Access Rate15%

How much of the important page set can AI crawlers actually reach.

Source: bot_requests + page_snapshots + audit_issues

Sentiment Score15%

Positive, neutral, and negative portrayal in generated answers.

Source: visibility_observations

Cross-Platform Coverage10%

How many of the four tracked engines cite the domain.

Source: visibility_observations

Source Authority10%

Strength of the domains surrounding your citations and mentions.

Source: visibility_observations

Tiers

Not Visible0-20

AI systems rarely surface or cite the domain.

Emerging21-40

The domain appears inconsistently and lacks repeatable coverage.

Established41-60

The domain is regularly cited and recognized in-category.

Authoritative61-80

The domain is consistently surfaced across the tracked engines.

Dominant81-100

The domain is first-to-mind in the active prompt set.

Update cadence

  • Prompt batches trigger a recalculation after new observations land.
  • Audit runs trigger recalculation after new access blockers or page states are found.
  • GA4 and GSC sync jobs can trigger recalculation when performance data changes.
  • Score history is stored for long-term trend analysis.

Caveats

  • The score is only as complete as the connected prompts, crawl data, and imports.
  • Estimated public scores for unclaimed domains use corpus evidence and are lower-confidence.
  • Methodology weights may evolve; changes are reflected in the changelog below.

Changelog

March 10, 2026 - v1 Renamed metric from m.score to AI Discovery Score. Published public methodology, wired score into the dashboard, public API, and AI Brand Checker surfaces.