Inbound Enquiry Triage Agent

Inbound enquiries can quickly become difficult to manage when they arrive through multiple channels and vary widely in quality, urgency, and intent. Sales and operations teams often spend valuable time reading, interpreting, and manually routing messages before any meaningful response can begin.

This project focused on creating an inbound enquiry triage agent that could automate the first layer of enquiry handling. The agent analysed incoming messages, identified the enquiry type, extracted important details, assessed urgency, and recommended or applied the correct routing path. This helped ensure that each enquiry reached the right destination with the right context.

The system was designed around real operational categories such as sales opportunities, support requests, partnership enquiries, supplier messages, spam, existing customer follow-ups, and urgent escalations. For each category, rules and AI classification logic were combined so the agent could make consistent decisions while still allowing human review when confidence was low.

A key feature was structured information extraction. Rather than simply labelling an email, the agent captured useful fields such as contact details, company name, requested service, budget signals, timeline, location, and next action. This information could then be pushed into CRM records, task queues, or notification workflows.

The result was a faster and more consistent inbound handling process. Teams spent less time sorting messages and more time acting on the right opportunities. The system also improved visibility by turning unstructured inbound communication into structured operational data that could be reported on and improved over time.