Agentic Trust IndexATI Screen v1.0 Beta

← Methodology

Methodology changelog

ATI is auditable itself.

Any change to scoring, authority levels, domains, gating, evidence maturity, question meaning or framework mappings is versioned and recorded here.

  1. v1.0-beta

    ATI Screen v1.0 Beta

    September 2026

    • Expanded ATI Screen from 12 to 20 questions, plus three system-context questions.
    • Introduced the five-domain model: Authority, Identity, Enforcement, Boundaries, Evidence.
    • Added authority boundary recommendations on a five-level scale (Observe → Expanded Autonomy).
    • Separated Authority Exposure from Trust Readiness and added explicit Evidence Confidence.
    • Added ten explicit, displayed gating conditions on the authority recommendation.
    • Replaced single evidence caps with the five-stage evidence maturity ladder.
    • Readiness is now the equal-weight mean of five domain scores (was 75% control strength + 25% evidence).
    • Added a versioned JSON export schema designed for future ATI Verify ingestion.

  2. v0.2-beta

    ATI Quick Assessment Beta v0.2

    2026

    • Added client-side PDF report generation.
    • Clarified that answers are not submitted, stored on a server, or emailed.

  3. v0.1

    ATI Quick Assessment prototype

    2026

    • Twelve questions sampling the Trust Architect's Handbook domains.
    • Readiness = 75% control strength + 25% evidence, with published score caps.
    • Three readiness bands: Unproven, Developing, Provisionally supported.