Proposal: AI Agent Interoperability for Institutions

Development Fund Proposal Submission

Proposal file: /proposals/proposal-agentic-ai-interoperability.md

---

Summary

Defines a Canton-native protocol enabling institutional AI agents to advertise capabilities, establish service relationships, and compose multi-party workflows, inheriting Canton's existing privacy, authorization, and atomicity guarantees. The protocol is implemented entirely at the Daml application layer through three interfaces (Agent Role Contract, Service Request, Service Record) and a two-phase request lifecycle, with adoption directly translating into Canton Coin burn through synchronizer traffic.

---

Checklist

• [x] Proposal file added under /proposals/
• [x] Milestones and funding amounts defined
• [x] Acceptance criteria included
• [x] Alignment with Canton priorities described

---

Notes for Reviewers

• Funding structure is performance-indexed. Milestones 1–3 are fixed (5M CC) and tied to verifiable on-chain and code-level deliverables. Milestone 4 (20M CC) releases in 1M CC tranches per $10M USD of cumulative attributable burn through the standard's contracts, capped at 20 tranches. An Adoption Bonus (25M CC at $1B burn, or 50M CC at $10B burn) sits on top.
• Backwards compatible. No changes to the Canton protocol, synchronization domain software, or existing Daml models. Implementable entirely at the application layer.
• Dependency: CIP-56 (Canton Coin fee denomination in Terms of Service).