VerificationPerformance Schema

Container: Performance.performanceAttributes Protocol Version: 2.0 Semantic Model: generalised Use Cases: Driving licence verification result, identity verification output, bank account verification confirmation, skill certificate validation, background check results Tag: verification performance result evidence audit

Overview

The VerificationPerformance schema captures the execution details and results of a single verification check. Each Performance unit in a contract corresponds to one verification type being executed — for example, one driving licence check or one identity check.

This is where the actual verification output lives: the result status, the verified data fields extracted from the authoritative source, processing steps for audit transparency, evidence artifacts, and confidence scoring.

Attachment Points

This schema extends beckn:performanceAttributes on the core Performance object.

Design Rationale

• verifiedData as open object: The shape of verified data varies significantly across

verification types. A driving licence returns holder name, vehicle classes, expiry date; an identity check returns name, date of birth, address. The open object pattern allows each verification type to return its specific fields without schema changes.

• processingSteps for transparency: Listing each processing step (input validation,

source query, data comparison) with timestamps enables audit trails and helps requesters understand where in the process a failure occurred.

• evidence for trust: Signed certificates, source response hashes, and API logs provide

cryptographic and documentary proof of the verification. This supports the use case's goal of portable, reusable trust.

• confidenceScore for fuzzy methods: When verification uses OCR, name matching, or

biometric comparison, the result may not be binary. The confidence score lets requesters apply their own thresholds.

• failureReason for diagnostics: Machine-readable failure codes enable automated retry

logic (e.g., retry on SOURCEUNAVAILABLE) and human escalation (e.g., alert on FRAUDSUSPECTED).

Non-Goals

• Subject PII management (→ VerificationContract)
• Pricing or payment for the verification (→ core Consideration / Settlement)
• Verification service description (→ VerificationResource)

Upstream Candidates

• VerificationResult with status enum — applicable to any compliance/trust domain
• ProcessingStep pattern — applicable to any multi-step service execution
• EvidenceArtifact pattern — applicable to any domain requiring audit trails
• confidenceScore — applicable to any fuzzy-matching or ML-based service