Orchestration is the fifth pillar of the Nomotic AMP. It handles how agents are scheduled, what triggers them to run, and how multi-agent workflows are coordinated.
Scheduling
From Orchestration → Scheduling, you can configure agents to run on a recurring schedule — hourly, daily, weekly, or on a custom cron expression. Scheduled runs are governed the same way as any other execution. Every action taken during a scheduled run is evaluated and recorded.
Triggers and webhooks
Triggers let you start an agent in response to an external event. From Orchestration → Triggers, you can configure:
Webhook triggers — an external system sends a POST request to your Nomotic webhook URL and the agent starts.
Event triggers — an agent starts in response to an internal Nomotic event, such as an escalation being approved or a drift alert firing.
Manual triggers — a team member starts the agent from the dashboard on demand.
Each trigger can be configured with conditions — the agent only starts if the incoming data meets specific criteria.
Execution control
From Orchestration → Execution Control, you can configure:
Priority — high-priority agents get processing preference during periods of high evaluation volume.
Concurrency limits — maximum number of simultaneous executions for an agent.
Queue behavior — what happens to queued executions when the limit is reached (wait or drop).
Timeout — maximum execution time before governance triggers an automatic interrupt.
Multi-agent workflows
For workflows involving multiple agents, Orchestration provides handoff coordination. You can define which agents can pass work to which other agents, and Nomotic governs the handoff — the receiving agent’s evaluation includes context about what the sending agent was doing. This prevents governance blind spots at agent boundaries.
MCP integration
Nomotic supports Model Context Protocol (MCP) for agents that use MCP-compatible tool servers. Governance wraps MCP tool calls the same way it