The project
Optician Jimmy Fairly faced growing numbers of order-tracking, return and repair requests. Each message required support staff to understand the request, retrieve the customer’s record and choose the right response. The aim was to reduce this work while retaining support rules and the brand’s voice.
Requests do not all fit the same categories and procedures change. The system had to follow the support team’s decision tree, rather than relying only on a classification learned from historical tickets.
The tool classifies the request, extracts relevant information and checks the customer record in the CRM. It applies support rules using replies approved by Jimmy Fairly’s teams.
60% of tickets are handled automatically and response time is down 40%. Automation covers request handling using customer records and the brand’s approved replies.
