Documentation
Trust Layer
Cryptographically verifiable consent, audit-ready receipts, and agent risk controls for agentic commerce.
Overview
The Trust Layer is the core of Clarify Systems' approach to safe agentic commerce. It provides cryptographic guarantees that every transaction was authorized, is auditable, and can be defended in disputes.
Consent Proof (Mandate Chain)
Every agent-initiated purchase is backed by a verifiable consent artifact that cryptographically binds user intent to cart contents to payment authorization.
How it works:
- 1 User confirms or pre-authorizes purchase intent through a secure confirmation flow
- 2 Consent is signed and timestamped, creating an immutable record
- 3 The mandate chain links: (1) user identity → (2) intent parameters → (3) cart snapshot → (4) payment authorization
- 4 Signatures use tamper-evident cryptography with replay protection (nonce + expiry)
- 5 Pre-authorized rules (for delegated tasks) specify spend limits, category restrictions, and validity periods
Revocation: Users can revoke pre-authorizations at any time. Revocations propagate immediately and invalidate any pending mandates.
Audit-ready Receipts
Every completed transaction produces a signed, immutable receipt with full provenance—ready for compliance reviews, financial audits, and dispute evidence.
Receipt contents:
Quote/cart snapshot: product SKUs, quantities, unit prices, tax, shipping, discounts applied
Consent chain references: links to the consent artifacts that authorized this transaction
Timestamps: request time, confirmation time, execution time, completion time
Payment references: PSP transaction ID, authorization code, settlement status
Execution references: merchant order ID, fulfillment tracking (when available)
Format: Receipts are available in JSON (machine-readable) and PDF (human-readable) formats. Both include cryptographic signatures for verification.
KYA (Know Your Agent)
Before executing any transaction, Clarify Systems verifies the agent's identity and enforces programmable policies based on risk signals.
Agent identity signals:
- Agent issuer (e.g., which LLM provider or application)
- Agent version and deployment context
- Agent status (active, suspended, rate-limited)
- Historical transaction patterns and anomaly scores
Programmable policies:
- Merchants can set per-agent or per-issuer limits (max transaction value, daily volume)
- Category restrictions (e.g., allow grocery, block electronics above threshold)
- Rate limits at agent, user, and merchant levels
- Risk signals are exportable to PSP or internal fraud systems
Dispute Evidence Pack
When a chargeback or dispute arises, Clarify Systems automatically generates a comprehensive evidence pack that can be submitted to payment processors.
Evidence pack includes:
- 1 Signed consent proof showing user authorized the transaction
- 2 Cart snapshot at time of consent (prices, quantities, totals)
- 3 Confirmation flow screenshots or recordings (if enabled)
- 4 Agent session logs showing the conversation context
- 5 Timestamp chain proving sequence of events
Note: The dispute pack is designed to meet card network evidence requirements. Success rates depend on dispute type and issuing bank policies.
Vendor neutrality
Clarify provides a consistent consent/receipt/risk layer regardless of the underlying model provider.
- Use your preferred agent platform (including Google/Amazon-based stacks) and keep your architecture flexible.
- Switch models or providers without changing your merchant adapter or trust layer integration.
- Same cryptographic guarantees and dispute-ready artifacts across all compatible agent runtimes.
Examples only. Clarify is not affiliated with or endorsed by Google, Amazon, or any other providers mentioned.
Merchant-ready kit
Get badge assets and ready-to-use copy to announce your integration with Clarify.
View merchant kit →