fleet-routing

LLM-based intent classification routes user requests by analyzing the input text and matching it to a structured routing table, dispatching the task to the appropriate Lil_Hawk specialist. This mechanism streamlines agent orchestration and reduces manual routing decisions for direct messages and escalations.

Automating user request routing is best handled by an LLM-driven classifier that evaluates incoming intents and dispatches them to the correct specialist endpoint. This approach supports structured routing tables for common mission types and includes retry and validation logic to ensure reliable delivery.

Yes, you can route ACHEEVY escalations to tactical execution by applying the LLM classifier to incoming requests. The system identifies the escalation intent and dispatches it to the chosen Lil_Hawk specialist's run endpoint with optional escalation support.

You need a structured routing table to map classified intents to their corresponding Lil_Hawk specialists. This table defines the dispatch targets for common mission types and ensures the LLM classifier routes requests to the correct run endpoint with proper validation.

If intent dispatch fails, the routing logic applies retry and validation mechanisms to ensure the request reaches the appropriate Lil_Hawk specialist. This validation logic confirms that the classified intent is correctly dispatched to the chosen run endpoint.