Beta NoticeThe AI Visibility Agency is currently in beta. Scores, rankings, methodologies, evidence models, platform functionality, and reporting outputs may change prior to general availability. Read beta terms

Methodology and Model Protection

v1.1.0 · Last updated 2026-06-15

Request Permission

Methodology Transparency and Intellectual Property

The AI Visibility Agency is committed to a clear and durable distinction between methodology transparency and methodology intellectual property. AIVA believes credible measurement requires public explanation of what is measured and why. AIVA also believes durable, defensible measurement requires that the precise mechanics of scoring remain proprietary.

Certain methodology details may be reviewed under confidentiality for procurement, audit, compliance, or enterprise due-diligence purposes. Such reviews may be conducted under a mutually agreed non-disclosure agreement.

Customers, partners, and analysts who require deeper methodological review for procurement, audit, or research purposes may request a methodology briefing under confidentiality through permissions@theaivisibilityagency.com.

AIVA discloses, and will continue to disclose:

  • The categories of evidence and signals considered in AIVA measurements.
  • The principles and design philosophy underlying AIVA scoring and ranking.
  • The evaluation dimensions assessed within each report or score.
  • Plain-language methodology summaries sufficient for customers, analysts, and auditors to understand what AIVA measures and why.

AIVA does not disclose, and treats as proprietary intellectual property:

  • Specific weightings applied to signals, categories, or evaluation dimensions.
  • Scoring formulas, normalization functions, and aggregation logic.
  • Calibration data, calibration logic, and threshold-setting procedures.
  • Ranking algorithms and tie-breaking logic.
  • Evidence models, prompt sets, and evidence-gathering procedures.
  • Confidence models, uncertainty quantification, and decision frameworks used to convert evidence into scores.

Methodology and Model Protection

AIVA's scoring methodologies, AI visibility models, ranking systems, evidence frameworks, weighting schemes, prompt sets, benchmarking approaches, evaluation rubrics, and underlying intelligence models (collectively, the "AIVA Methodologies") are proprietary and confidential to The AI Visibility Agency.

AIVA publishes plain-language explanations of its methodologies to support transparency, customer due diligence, and analyst review. These public explanations are provided for informational purposes only and do not convey any license to the underlying methodology, models, data structures, calibration data, or implementation.

The AIVA Methodologies may not be reverse engineered, reconstructed, replicated, benchmarked against for the purpose of producing a competing product, or incorporated into third-party scoring systems, indices, or AI evaluation services without a written license from The AI Visibility Agency.