GKS-5 introduces a five-layer reference architecture for governed enterprise AI knowledge systems. It focuses on ingestion, storage and semantic indexing, retrieval, model orchestration, application integration, and cross-layer controls such as metadata, authorization, freshness, observability, evaluation, and versioning.
AGL-1 introduces a vendor-neutral control-plane model for governing retrieval, memory, policy, provenance, observability, and agentic execution in enterprise AI systems.
CAGE-1 introduces an evaluation framework for assessing whether enterprise AI agents are ready for deployment across authority, policy enforcement, retrieval trust, memory integrity, tool safety, auditability, human oversight, conflict handling, safe failure, operational readiness, business fitness, and Prebind Assurance.