Stage 4 - Consensus and Finalization

After zero-knowledge proofs are generated for individual atomic claims, the protocol combines them into a final proof for the original composite claim. This combination process ensures that the final proof maintains all privacy properties of individual atomic proofs while demonstrating the validity of the complete claim.

The consensus mechanism at this stage serves multiple purposes. It aggregates atomic claim proofs into composite claim proofs, validates that all required atomic claims were successfully verified, applies any additional verification requirements specified in the original claim, and generates the final proof that will be delivered to the claimant.

For high-stakes claims or ambiguous cases, this stage may include human verification oversight. Designated human verifiers review the evidence, agent verification results, and proposed proofs to ensure accuracy. Human verifiers are particularly valuable for claims that require judgment, interpretation, or domain expertise that AI agents may lack. The protocol implements economic incentives that reward accurate human verification and penalize errors, ensuring high-quality human oversight.

Once consensus is reached and any required human verification is complete, the protocol finalizes the verification. Finalization involves recording the verification result on-chain for transparency and immutability, generating the final zero-knowledge proof with all required properties, notifying the claimant that verification is complete, and making the proof available for retrieval and use.