TrustGate Decision Methodology
Version 1.0
Published February 2026 · TrustGate by Exonara Technologies
This document describes the high-level methodology used by TrustGate to assess datasets for decision-readiness and to produce verdicts and Evidence Packs. It is intended for auditors, compliance officers, and legal teams who need to verify how an assessment was performed. The exact implementation (weights, thresholds, and algorithms) is proprietary.
Trust principle. Users are meant to trust TrustGate for high-stakes decisions; any mistake can be costly. We therefore err on the side of caution: we only issue Approve when all assessed dimensions (privacy, quality, bias) meet the bar for the decision context and there are no duplicate rows or severe breaches. When in doubt we issue Conditional or Block and explain why, and we provide a path to Approve so every issue can be remediated.
1. What we assess
TrustGate evaluates three dimensions before a high-stakes, data-driven decision:
- Privacy (PII and sensitivity): Whether the dataset contains personally identifiable information or other sensitive data that would create regulatory or reputational risk if used without appropriate controls.
- Data quality: Completeness, duplicates, outliers, and consistency. Poor quality can lead to biased or incorrect decisions.
- Bias and fairness: Demographic and statistical imbalance that could lead to discriminatory or unfair outcomes (e.g. in hiring, lending, or allocation).
2. How we derive the verdict
Assessments are performed in the context of the decision type (e.g. hiring, lending, medical, operational) and impact level (low, medium, high). TrustGate applies context-aware rules and thresholds to produce a single governance verdict:
- Approve — The data, as assessed, is suitable for use for this decision within the stated context.
- Conditional — Use is acceptable subject to documented conditions or remediation (e.g. applying certain fixes or controls).
- Block — The data is not suitable for this decision in its current form; remediation is required before use.
A decision score (0–100) and a confidence score from the assessment engine are provided to support the verdict. The exact formulas and thresholds are proprietary and may be updated in future methodology versions.
3. Evidence Pack
When requested (and where permitted by your plan), TrustGate generates an Evidence Pack that includes:
- Attestation that the assessment was performed on a given date using this methodology (TrustGate Decision Methodology v1.0).
- Verdict, decision score, and a high-level summary of findings.
- Mapping to relevant compliance frameworks (e.g. GDPR, CCPA, EU AI Act, fair lending) where applicable.
The Evidence Pack is intended for internal audit and compliance and for sharing with auditors or regulators as evidence of due diligence. It does not constitute legal advice.
4. Versioning
This methodology is versioned. Assessments performed under v1.0 remain valid under v1.0. Future versions (e.g. v1.1, v2.0) will be published and dated; the version used for each assessment is recorded in the Evidence Pack and in the product.
Attestation statement. Every Evidence Pack states: “TrustGate attests that this assessment was performed on [date] using TrustGate Decision Methodology v1.0.” This allows auditors to verify the methodology used.
5. Contact
For questions about the methodology or to request a copy for your records: preciousoladapo@exonara.net. TrustGate by Exonara Technologies.
← Back to TrustGate · The Standard · Terms of Service