Minimum control baselines for enterprise AI. Use these pages for practical controls, required records, and operating handoffs.
Minimum controls before moving beyond experimentation.
Identity, entitlement, least privilege, revocation.
Open standardRuntime gateway controls for model and tool traffic.
Open standardModel approval, routing, monitoring, retirement.
Agent controls for autonomy, tooling, escalation, and telemetry.
Open standardControls for AI-callable tools/APIs/workflows.
Open standardControls for MCP server registration, security, and operation.
Open standardData classification, restrictions, and retention for AI use.
Open standardPrompt safety and output release controls.
Open standardRAG, ontology, graph, provenance, grounding quality.
Open standardLifecycle telemetry and audit evidence requirements.
Open standardPre/post-production test baselines and release gates.
Open standard