ChinaFinance and insurance2025 – ongoingChina Everbright Bank
Customer-service summaries and case routing
China Everbright Bank reported cutting agent form-filling time by 13 seconds with AI summaries
Problem
Agents needed to review earlier customer interactions and capture key details from lengthy calls. Omitted information could weaken service records.
Approach
The bank embedded summaries of previous calls and service tickets in its customer-service system. A fine-tuned LLM generated conversation summaries, while smaller models classified tickets for routing under predefined rules.
Results
Takeaway
The bank used language generation for summaries and smaller classification models for ticket handling.
Categories
- Size
- Large enterprise
- Industry
- Finance and insurance
- Business function
- Customer service
- Innovation type
- Process innovation
- AI technology
- Generative AIPredictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The bank described routine use of customer-history summaries and automated ticket classification. It attributed a 13-second reduction in form-filling time to its conversation-summary model.
Yang Bingbing described deployed agent assistance and ticket summaries. He reported shorter overall call handling, a different measure from form filling.
Full details are in the original sources.