AI GOVERNANCE / PUBLIC WALKTHROUGH
NeuralOps Control Centre
Synthetic sample insights. Execution and registry changes are disabled in the public walkthrough.
AI simulation: The commands below execute read-only, server-side rules. No external LLM, autonomous clinical decision or generative patient access is connected.
Sample agent tools
RULES MODENo login required. The public demo is read-only; staff login is required to run tools.
Reception Agent
{"upcoming_appointments":4}
Inventory Agent
{"low_stock_items":1}
Executive Agent
{"active_patients":6,"branches":3}
Execution principles
✓Role and branch scope enforced by backend
✓Agent-specific tool allowlist
✓Read-only SQL business insight tools
✓Approval decisions logged with actor identity
✓No prompt-to-SQL or unrestricted tool execution
✓Smart Routing tiers are configuration placeholders
