Legal & Compliance
Immutable audit receipts, fairness gates, and zero-knowledge proof of compliance
Legal and regulatory workflows require provable fairness, immutable audit trails, and privacy-preserving verification. PrimeSwarm captures every agent decision with cryptographic receipts, fairness gates detect bias using the Four-Fifths Rule, and OnlyState enables zero-knowledge proofs of group membership without exposing individual identities.
Regulated industry pathway
Legal & Compliance is a regulated industry. PrimeSwarm deploys in four phases: Discovery, Build, Deploy, Maintain. We configure the governance stack, write the ONLY Lang scripts, and run the DGV test card suite. You approve every script before it goes live.
Discovery
Week 0. We map your regulatory constraints, infrastructure, and decision workflows.
Build
Weeks 1-4. We configure the gate, write ONLY Lang scripts, and run DGV test cards. You sign off on every script.
Deploy
Week 4-5. We deploy to your VPC or on-prem. You receive a signed verification report.
Maintain
Ongoing. Quarterly reviews, DGV re-verification, script updates. You retain full control.
Why this matters
Counsel cannot defend “the model said so.” They need a refusal that happened before the act, a receipt of state, and a map from claim to regulation.
Safety and legal compliance
Pre-effect refusal and Four-Fifths checks on high-risk decisions. TPNN is the injection defense for legal assistants. DGV is the evidence harness. We do not replace a law firm or a GRC suite.
- EU AI Act (high-risk mapping). DGV test cards and TDO / receipt packets
- NIST AI RMF. Govern / Map / Measure language via DGV + PIR gap
- ISO 42001 (mapping). Claim registry and reproducible evidence
- Employment / credit fairness. Four-Fifths Rule in PrimeSwarm
Start this month
- Estate Protect: Node Exposure of Harvey / Copilot / Claude writes
- HITL on file / send — unresolved cite cannot GREEN
- DGV verifier bundle and evidence receipts
- Credit Advisor / hiring-gate pattern
Not yet — we will not sell this
- · A production caselaw search product (TPNN legal engine is a prototype)
- · A Harvey / Lexis / Westlaw copilot replacement
- · Outsourced legal advice
How Our Products Apply
4 products with specific use cases in Legal & Compliance
PrimeSwarm
View product →Use Case
Deploy compliance-review agents that evaluate loan applications, hiring decisions, and insurance claims with automatic Adverse Impact Ratio calculation. Every decision produces a cryptographic receipt stored in HelixDB.
Key Benefits
- Automatic Adverse Impact Ratio calculation using the Four-Fifths Rule
- Pre-Effect Refusal — biased decisions are blocked before execution, not flagged after
- Cryptographic audit receipts capture exact agent state at decision time
- SIEM integration for real-time compliance monitoring in Splunk/Datadog
- JWT authentication with RBAC for role-based access to sensitive workflows
OnlyState
View product →Use Case
Prove compliance with regulatory group membership requirements without exposing individual identities. ZK-PIR circuits verify membership against a constant-size on-chain state — no databases, no identity arrays.
Key Benefits
- Zero-knowledge proofs of membership without revealing individual identities
- O(1) on-chain state — regulatory groups of any size stored as one 16-element array
- Groth16 proofs verifiable by any third party without trusted setup per verification
- Immutable on-chain audit trail of group state changes
TPNN
View product →Use Case
Protect legal AI assistants from prompt injection that could alter legal advice. TPNN's topological constraints ensure that adversarial inputs cannot redirect the model's output outside the valid legal reasoning manifold.
Key Benefits
- Mathematical guarantee against prompt injection in legal document analysis
- Output constrained to valid legal reasoning manifold
- No probabilistic false negatives — defense is structural, not heuristic
DGV Verifier
View product →Use Case
Produce a reproducible evidence packet for an AI-use policy: test cards, SHA-256 receipts, and a claim registry mapped to EU AI Act, NIST AI RMF, and ISO 42001 language.
Key Benefits
- Self-verification you can hand to counsel — not a rented SOC 2 logo
- Maps claims to regulation; does not impersonate a notified-body certificate
- Replayable receipts for “what was tested, when”
Other Industries
All industriesBanking & Markets
Regime detection, fair-lending gates, and audit-ready agent governance
Healthcare
PII-safe agents, adversarial defense, and mastery-tracked medical education
Education & Research
Mastery-tracked learning, autonomous research generation, and knowledge graphs
Sales & CRM
Agentic inbox reading, company enrichment, and workflow automation