Scoring methodology

The AI Neutrality Score

Every tool in the directory is scored using the ANS — a four-dimension framework designed to measure how trustworthy, transparent, and rigorous a compliance tool actually is.

How the score works

The ANS is a 0–100 composite score. Each of the four dimensions is scored independently and weighted by importance. A tool scoring above 85 is considered best-in-class. Scores between 60–79 indicate a solid tool with meaningful gaps. Scores below 60 indicate significant concerns that practitioners should investigate before purchasing.

30
points
Transparency
25
points
Neutrality
30
points
Methodology
15
points
Evidence Quality
30
pts

Transparency

Does the vendor clearly disclose how their product works, how evidence is collected, and how audits are conducted?

High-scoring tools publish their methodology publicly, maintain clear audit trails, provide exportable evidence packages, and avoid black-box automation that auditors cannot independently verify.

Scoring bands

25–30
Full transparency
Public methodology docs, exportable audit packages, versioned changelog, open API for evidence retrieval.
15–24
Good transparency
Documentation available on request, audit exports exist but require manual steps, changelog maintained internally.
5–14
Partial transparency
High-level docs only, limited export options, no public changelog. Evidence collection logic not disclosed.
0–4
Opaque
Black-box automation, no documentation of evidence logic, auditors must trust vendor outputs without verification.

Signals we look for

  • Public documentation of evidence collection logic
  • Audit-ready export formats (PDF, CSV, ZIP)
  • Changelog and update history published
  • API access for third-party evidence retrieval
  • Clear disclosure of what is automated vs. manual
25
pts

Neutrality

Does the vendor have conflicts of interest that could bias their assessments?

We evaluate vendor relationships with auditors, reseller arrangements, and whether compliance recommendations are influenced by commercial incentives rather than objective control coverage.

Scoring bands

21–25
Fully independent
No auditor partnerships, no framework exclusivity deals, pricing not tied to compliance outcomes, recommendations are framework-agnostic.
13–20
Mostly independent
Preferred auditor relationships disclosed, minor commercial arrangements that do not materially bias recommendations.
5–12
Some conflicts
Exclusive auditor partnerships, framework coverage skewed toward commercially advantageous certifications.
0–4
Significant conflicts
Revenue-sharing with auditors, certifications bundled with paid add-ons, recommendations demonstrably biased by commercial interests.

Signals we look for

  • No exclusive auditor partnerships
  • Framework coverage not tied to revenue deals
  • Independent third-party validation of scoring
  • Conflict-of-interest policy published
  • No reseller arrangements with auditing bodies
30
pts

Methodology

How rigorous is the technical approach to compliance automation?

We evaluate coverage depth across frameworks, quality of control mapping, evidence validation logic, and how the platform handles ambiguous or overlapping controls across multiple frameworks.

Scoring bands

25–30
Rigorous
Deep control mapping with citations, cross-framework overlap handled explicitly, evidence validation logic documented and testable.
15–24
Solid
Good coverage of major frameworks, control mapping documented, some gaps in edge cases or newer frameworks.
5–14
Basic
Surface-level control mapping, limited framework coverage, evidence collection relies heavily on self-attestation.
0–4
Weak
Checklist-based approach with no meaningful control validation, framework coverage incomplete or outdated.

Signals we look for

  • Control mapping with explicit framework citations
  • Cross-framework gap analysis capability
  • Evidence validation beyond self-attestation
  • Regular framework update cadence documented
  • Handles overlapping controls across frameworks
15
pts

Evidence Quality

How strong and verifiable is the evidence the tool produces for auditors?

Auditors need evidence that is timestamped, tamper-evident, and traceable to specific controls. We evaluate whether the platform produces audit-grade artifacts or merely screenshots and self-reported data.

Scoring bands

13–15
Audit-grade
Tamper-evident logs, cryptographic timestamps, direct API integrations with source systems, auditor portal with read-only access.
8–12
Good
Timestamped exports, integration-sourced evidence for major controls, some manual steps required for edge cases.
3–7
Adequate
Evidence collected but relies on screenshots or self-attestation for significant portions of controls.
0–2
Weak
Primarily self-reported, no tamper-evidence, auditors cannot independently verify source of evidence.

Signals we look for

  • Cryptographic or tamper-evident audit logs
  • Direct integration with source systems (AWS, GitHub, etc.)
  • Read-only auditor portal
  • Evidence linked to specific control requirements
  • Automated evidence refresh cadence

Update cadence

Scores are reviewed quarterly. When a vendor ships a material update — new integrations, pricing changes, or methodology shifts — we re-evaluate within 30 days. Score changes are logged in the tool’s changelog so practitioners can track how a vendor has evolved.

Vendors may submit evidence for re-evaluation at any time by emailing hello@compliancedirectory.io. Evidence submissions are evaluated within 14 business days.

Independence policy

No vendor pays to influence their ANS score, their position in the directory, or any editorial content about their tool. Commercial relationships (Pro listings, Enterprise contracts) are strictly limited to listing visibility features — they have zero effect on scoring.

Scorers are required to disclose any prior relationship with a vendor before conducting a review. Tools scored by a reviewer with a conflict of interest are independently verified by a second reviewer.