Skip to content
TruthVouch Docs

Understanding Your Score

TruthVouch assessments automatically calculate dimension-based scores that measure your organization’s AI maturity across governance, technical, compliance, and operational areas. Learn how scores are calculated and what they mean.

Maturity Dashboard

Overall Score

Your Overall AI Maturity Score is a weighted average of five core dimensions:

DimensionWhat It Measures
MonitoringHow well you observe and alert on AI outputs, detection capabilities, observability
ComplianceRegulatory posture, policies, audits, compliance training completion
GovernancePolicy discipline, configuration controls, decision boundaries, agent autonomy
TransparencyModel cards, explainability artifacts, trust center publication, audit trails
OperationsIncident response, supply-chain mapping, DR drill cadence, team readiness

Each dimension is scored on a scale of 0–5, and the overall maturity level is the weighted average.

Dimension Scoring

Each dimension measures specific aspects of your AI maturity and is scored 0–5:

1. Monitoring (0–5)

How well you observe and alert on AI outputs:

  • Real-time AI output scanning (hallucination detection, PII detection, anomaly detection)
  • Alert infrastructure (routing, escalation, dashboards)
  • Observability and logging (centralized logs, query traceability, response attribution)
  • Integration with AI systems (SDKs, APIs, manual review processes)

Score 0: No monitoring; manual log collection only Score 5: Continuous automated detection with real-time alerts and centralized observability

2. Compliance (0–5)

Regulatory posture and compliance readiness:

  • Industry compliance frameworks mapped (GDPR, SOX, HIPAA, EU AI Act, ISO 42001)
  • Compliance checklist completion and gap tracking
  • Audit preparation and evidence collection
  • Compliance training program completion

Score 0: No formal compliance posture; unaware of applicable frameworks Score 5: All frameworks mapped, gaps closed, audit-ready documentation, training complete

3. Governance (0–5)

Policy discipline and control enforcement:

  • Formal AI governance policy (documented, approved, communicated)
  • Policy evaluation against AI decisions (manual or automated)
  • Configuration controls for agent autonomy and action boundaries
  • Risk assessment and approval workflows for AI deployments

Score 0: No formal policy; ad-hoc decision-making Score 5: Comprehensive policies, continuous evaluation, clear autonomy boundaries, approval workflows

4. Transparency (0–5)

Explainability and public trust:

  • Model cards and technical documentation
  • Explainability artifacts (decision trails, source attribution)
  • Trust center publication and public verification pages
  • Cryptographic attestation and certification

Score 0: No documentation; no public trust mechanisms Score 5: Comprehensive model cards, explainability artifacts, published trust center, cryptographic verification

5. Operations (0–5)

Operational readiness and incident response:

  • Incident response procedures (classification, escalation, root cause analysis)
  • Supply-chain mapping and risk assessment (vendors, models, data sources)
  • Disaster recovery (backup, recovery testing, failover procedures)
  • Team readiness (skills, training, on-call coverage)

Score 0: No formal incident response; team gaps; no DR plan Score 5: Documented procedures, regular testing, trained team, comprehensive supply-chain visibility

Score Interpretation

Maturity Levels (0–5 scale)

Score RangeLevelWhat It Means
4.0–5.0AdvancedBest-in-class; comprehensive controls across all dimensions
3.0–3.9DevelopedStrong foundation; some gaps remain in specific areas
2.0–2.9ManagedBasic processes; multiple areas need strengthening
1.0–1.9InitialEarly stage; foundational work required
0–0.9Ad-hocLittle to no formal governance

Red Flags (Score < 2)

Scores below 2 in any dimension indicate:

  • Monitoring < 2: No real-time detection; relying on manual processes
  • Compliance < 2: Regulatory risk; unprepared for audits
  • Governance < 2: No formal AI policy; uncontrolled AI deployments
  • Transparency < 2: No audit trails or explainability; public trust issues
  • Operations < 2: Team gaps; no formal incident response plan

Benchmarking

Your scores are compared against:

  • Industry peers (Finance, Healthcare, Tech, etc.)
  • Company size (1-10, 11-50, 51-250, 251-1000, 1000+)
  • Geography (US, Europe, APAC)

See Benchmarks to understand your percentile ranking.

Scoring Methodology

Assessment Approach

TruthVouch uses two scoring methods:

  1. LLM-Based Scoring (Primary): An LLM evaluates your usage patterns and practices to generate rich narratives and nuanced scores across the five dimensions
  2. Deterministic Fallback: If LLM scoring unavailable, a heuristic scoring engine evaluates your documented practices based on assessment answers

Dimension Scoring

Each dimension is scored on a continuous scale of 0–5:

  • 0–1: Minimal or no formal practices in this dimension
  • 1–2: Basic awareness; some foundational elements in place
  • 2–3: Established practices; most key controls present
  • 3–4: Mature implementation; strong across most areas
  • 4–5: Advanced, best-in-class implementation

Partial Credit

The assessment allows for granular scoring:

  • Full score (5.0): All aspects mature and continuously improved
  • High score (4.0–4.9): Strong across dimension with minor gaps
  • Moderate score (3.0–3.9): Solid foundation with some targeted improvements needed
  • Low score (1.0–2.9): Basic practices with significant gaps
  • Minimal score (<1.0): Little to no formal practices

Confidence Scoring

The assessment includes a Confidence Score (0-100) for the overall maturity level:

  • 90+: High confidence; comprehensive information; stable assessment
  • 70-89: Moderate confidence; minor clarifications could refine score
  • 50-69: Lower confidence; recommend detailed audit for accuracy
  • <50: Low confidence; suggest external validation or consultant review

Low confidence indicates you may want additional assessment (interviews, audit) before major decisions.

Score Stability

Assessment scores are relatively stable but can fluctuate:

Month-to-Month Variation

  • ±0.5 points: Natural variance from assessment interpretation
  • ±1.0 point: Significant change; investigate what changed

Seasonal Patterns

  • Scores may dip after compliance events (new regulations)
  • Improve after training programs are completed
  • Vary with team turnover or staffing changes

Improvement Tracking

Historical Comparison

After your second assessment (minimum 30 days later):

  • See dimension-by-dimension progress
  • Understand which improvements moved the needle
  • Track overall maturity trajectory

Projected Trajectory

Based on your improvement rate:

  • “At current pace, you’ll reach 4.5 Governance by Q3 2026”
  • “You’ve plateaued at 3.0 Monitoring — likely need external help”
  • “Fastest improvement: Operations (+0.5 points/quarter)“

Customization (Enterprise)

Enterprise customers can:

  • Reweight dimensions (e.g., Compliance = 40% instead of 25%)
  • Add custom questions (company-specific concerns)
  • Adjust scoring rubrics (align with internal standards)

Contact your CSM to customize scoring.

Next Steps

  1. Review your dimension scores — which area is weakest?
  2. Compare to benchmarks — where are you vs. peers?
  3. Read improvement recommendations — prioritize the top 3 gaps
  4. Plan actions — assign owners and timelines
  5. Re-assess in 90 days — measure progress