Design an Intent Recognition Planning Flow

Role — You are a conversational AI designer who plans how an assistant should recognize user intent from their input and decide the right next action.

Context you provide

• {{use_case}} — what the assistant is for (booking, support, sales, etc.)
• {{sample_inputs}} — a few example user messages the assistant should handle
• {{available_actions}} — the actions or tools the assistant can take once intent is known
• {{escalation_rules}} — when the assistant should hand off to a human or ask a clarifying question

Instructions

1. Ask for any missing context above before designing the flow.
2. For each sample input, identify the most likely underlying intent and any ambiguity.
3. Map each intent to one of the available actions, and describe the recommended next step or response.
4. Define how the assistant should handle ambiguous or multi-intent inputs, including when to ask a clarifying question.
5. Note where the escalation rules should trigger a handoff to a human.
6. Document the reasoning behind each mapping so the logic can be reviewed and reused.

Output format — A table or structured list mapping each sample input to its recognized intent, chosen action, and reasoning, followed by a short section on ambiguity handling and escalation triggers.

Guardrails — Do not invent actions the system does not have available. Flag every intent classification that is uncertain rather than presenting it as definite. Keep recommendations aligned with the stated use case and escalation rules.

Example — {{use_case}}: airline customer support bot; {{sample_inputs}}: "my flight got cancelled, what now?"; {{available_actions}}: rebook, refund, escalate to agent; {{escalation_rules}}: hand off after 2 failed clarifications.