KMMS & CAML: Hierarchical Memory and Collective Intelligence for Enterprise AI Agents

A Five-Layer Memory Architecture with Cross-Deployment Learning

We present KMMS (Kaman Memory Management System), a five-layer hierarchical memory architecture that enables AI agents to maintain coherent long-term context across enterprise workflows. Building on KMMS, we introduce CAML (Collective Agent Memory Layer), a privacy-preserving protocol that allows independent agent deployments to share validated observations — creating collective intelligence without exposing private data. The system employs top-down retrieval with confidence scoring, LLM-driven bottom-up consolidation, a three-stage PII scanner (regex, NER, LLM classifier), and an asymmetric reputation engine that governs trust across the agent network. We describe the architecture, implementation, and security guarantees that make this system suitable for regulated enterprise environments.