Kiko: New programming model lets agents enact interaction protocols
Researchers unveil Kiko to let agents follow protocols without communication boilerplate
A team of researchers from academia has introduced Kiko, a programming model designed to simplify building multiagent systems by focusing on interaction protocols. Kiko lets developers write one or more 'decision makers' per agent, each of which chooses from a set of valid decisions and ensures messages are mutually compatible. By completely abstracting away the communication layer, Kiko allows developers to focus on business logic rather than low-level messaging details. The model supports common decision-making patterns, making it practical for real-world decentralized systems.
The Kiko paper provides operational semantics for the model and formally establishes that agents built with it are protocol compliant and can enact any protocol. This means developers get provable correctness guarantees for agent interactions while maintaining flexibility in internal decision logic. Kiko addresses a key gap in agent programming: bridging internal and public decisions without coupling them to specific communication services. The work has implications for industries relying on multiagent coordination, such as supply chain management, autonomous vehicles, and distributed AI services.
- Kiko's decision makers automatically select from valid, mutually compatible message decisions.
- The model completely abstracts away underlying communication services.
- Formally proven to ensure protocol compliance and support any protocol enactment.
Why It Matters
Kiko simplifies decentralized multiagent development by letting teams focus on business logic over communication plumbing.