How might we use AI to help service desk analysts classify, prioritize and resolve incoming requests faster while keeping humans in control?
Learn more about the Challenge
Q&A Session by Swiss Life Thu 15 Sep at 15.00h
👉 signup here
Details/materials on GitHub
Problem
Support desks across Swiss Life spend significant time manually triaging requests, searching fragmented documentation and rewriting similar answers. Resolution speed and quality depend heavily on the individual analyst handling the ticket, resulting in inefficiencies and inconsistencies.
Objective
Build an AI agent that ingests an incoming email or ticket, classifies and prioritizes it, retrieves relevant knowledge and historical resolutions, and drafts a suggested response for analyst review. The analyst remains in control and approves, edits or rejects the recommendation.
Support for Hackers
- Sample anonymized or mocked AM Service Desk tickets
- Example knowledge articles and procedures
- Domain owner support
- Business process guidance
Technical Preferences
Azure AI, Azure OpenAI, Microsoft Copilot Studio, RAG architectures, Semantic Kernel and workflow orchestration tools. Open to any suitable stack.
Why hack? A real operational problem with measurable outcomes such as time-to-resolution, classification accuracy and draft acceptance rate. Can be prototyped in two days using mocked ticket data and scaled across AM Service Desk, IT support, client service and other Swiss Life functions.
About the Challenge Partner
Swiss Life empowers people to lead financially self-determined lives and is committed to responsible, innovative AI in insurance and pensions.
As a Hackathon Partner of Swiss {ai} Weeks, we support collaborative innovation to turn AI research into real-world solutions for society, business, and individuals. By integrating secure and transparent AI, we enhance customer experience, drive operational excellence, and help shape the future of the financial sector in Switzerland and beyond.