Trust & Compliance

Last updated: March 2026

TrancheBook is built on principles of transparency, data integrity, and responsible AI use. This page describes our compliance framework, governance practices, and the measures we take to earn and maintain your trust.

1. Information Platform Model

TrancheBook operates as an information and intelligence platform for private markets. It is important to understand what we are and what we are not:

Where users express interest in transactions through our IOI facilitation feature, we act solely as a communication channel between qualified parties. We do not execute, settle, or guarantee any transaction.

2. KYC and AML Compliance

For users who wish to access IOI facilitation features, we implement Know Your Customer (KYC) and Anti-Money Laundering (AML) procedures:

• Identity verification: Government-issued identification is required for IOI access. Verification is performed through our compliance workflow.
• Accreditation checks: Users accessing IOI features must demonstrate that they meet applicable professional or accredited investor criteria.
• Sanctions screening: All users with IOI access are screened against applicable sanctions lists, including OFAC, EU, and UN sanctions lists.
• Ongoing monitoring: We perform periodic re-verification and ongoing screening for users with IOI access.
• Suspicious activity: We maintain procedures for identifying and reporting suspicious activity to relevant authorities as required by law.

3. Eligibility Gating

The Platform applies tiered access controls to ensure appropriate content reaches appropriate audiences:

4. Audit Trail

TrancheBook maintains comprehensive audit trails for compliance and transparency:

• AI analysis logs: Every analysis generation is logged with the input data, model version, and output. This allows us to reproduce and review any analysis.
• Score history: Historical scores for each company are preserved, allowing users and regulators to track how assessments have changed over time.
• IOI records: All IOI submissions, modifications, and withdrawals are recorded with timestamps and user identification.
• Access logs: User access to restricted content and features is logged for compliance purposes.
• Data change logs: Changes to company data, user accounts, and system configurations are tracked with full change history.

5. Data Security

We implement robust security measures to protect your data:

• AES-256 encryption at rest for all sensitive data, including portfolio information.
• TLS 1.3 encryption for all data in transit.
• Per-user logical data isolation ensuring no data leakage between accounts.
• Regular security assessments and penetration testing.
• Access controls based on the principle of least privilege.
• Automated monitoring for unauthorised access attempts.

For detailed information about our security practices, see our Security page.

6. AI Governance

We maintain a structured AI governance framework to ensure responsible use of artificial intelligence:

6.1 Model Versioning

Every AI model and prompt version used by Cole is tracked and versioned. When we update models or analysis methodologies, we maintain records of previous versions and can reproduce historical analyses. Users are notified of material changes to analysis methodology.

6.2 Output Logging

All AI-generated outputs are logged with their inputs, model version, and generation parameters. This provides a complete audit trail for any piece of analysis and supports reproducibility and review.

6.3 Bias Monitoring

We actively monitor Cole's outputs for systematic biases across dimensions including:

• Sector bias: whether certain sectors receive systematically higher or lower scores.
• Geography bias: whether companies in certain regions are evaluated differently.
• Stage bias: whether early-stage or late-stage companies are systematically favoured.
• Data availability bias: whether companies with more public data receive higher scores regardless of quality.

When biases are identified, we adjust our models and methodologies to mitigate them. Significant bias findings are disclosed in our AI Disclosures page.

7. Data Quality Governance

Our Data Quality Level (DQL) framework is central to our commitment to transparency:

• Every analysis output displays its DQL rating (1-5), providing users with immediate visibility into data reliability.
• DQL ratings are assigned algorithmically based on source type, recency, cross-validation, and completeness.
• We do not generate IC packages for companies where data quality is insufficient to support meaningful analysis.
• Data quality metrics are tracked over time and reported internally for continuous improvement.
• Users can view the data sources underlying any analysis to independently assess data quality.

Questions

For questions about our compliance framework, contact us at compliance@tranchebook.com.