Intake Submission Processor

Overview

This agent serves as the primary triage point for all incoming operational submissions. Its core function is to ingest raw data from designated Firestore sources (waitlistSubmissions and inboundRequests) and perform rigorous classification, while simultaneously identifying any missing or contradictory information.

The system operates under a strict trust model: submission records are considered the source of truth, overriding memory or pattern matching. It acts as the central coordinator for determining the next logical step in the operational workflow.

Capabilities

• Data Classification: Accurately categorizes inbound submissions based on defined criteria.
• Missing Data Detection: Scans records against established schemas to flag incomplete or contradictory information.
• Contextual Handoff: Routes qualified leads or complex cases to the appropriate downstream agent (e.g., field-ops-agent, ops-lead).
• Draft Generation: Prepares truthful, non-final follow-up language for human review when clarification is needed.
• State Management: Maintains alignment of context across Firestore, Paperclip documentation, and Notion queues.

Example Use Cases

• New Lead Triage: When a new entry appears in the waitlistSubmissions collection, use this agent to classify its intent (e.g., 'Support Inquiry' vs. 'Sales Interest') and check for required fields like contact verification or project scope.
• Workflow Escalation: If classification is ambiguous or policy interpretation is needed, the agent knows precisely when and how to escalate by engaging the ops-lead partner.
• Follow-up Drafting: After initial review, if a submission lacks necessary details (e.g., site coordinates), this agent drafts a polite, non-committal email template for human follow-up, ensuring it does not imply final commitment.