Classifying Inbound Customer Requests
Sort inbound requests into your own categories, pull out what each one needs, and separate the ones that are really complaints.
About this use case
Inbound requests arrive as free text and get categorized by whoever opens them. The categories drift, the urgent ones sit in the general queue, and the ones that are actually complaints get handled as service requests — which is the mistake with the longest tail.
This use case classifies each request against your taxonomy, extracts what the request contains, and flags anything that meets your complaint definition for separate handling.
A batch of emails comes back classified with what each asks for and what it is missing; a payoff request that also carries language meeting the complaint definition is classified both ways and the complaint routed separately; a message asking three unrelated things is separated into three.
Documents to consider
- Your request category taxonomy and definitions
- Routing rules and service levels by category
- Complaint definition and identification criteria
- Required information by request type
- Examples of correctly categorized requests
The prompt
Written to be pasted straight into your AI assistant, with the documents it asks for attached. Request access below and we’ll send it over.
Files
Nothing to download here. This prompt runs against your own institution’s material — the documents it lists above.
More in Retail & Branch Banking
Customer Support
Give a support agent a customer-ready answer drawn from the knowledge base, fee schedule and product terms, with the decision path documented.
Client Advisement
Draft a client-ready explanation from approved product material and compliance-cleared scripts, with the required disclosures already in place.
Cross-Sell and Upsell Opportunities
Match a customer's holdings and behavior against eligibility rules and current offers, then draft outreach that is relevant rather than generic.