1) Pick a Scenario
Interviewer picks one of 10 fictional scenarios. The next page will simulate the agent console, pre-filling likely “reason for call”.
2) What we're proving
- Reality: 20M subs, 5k agents; AHT spikes mid and end of month.
- Pain: AHT spikes around payment cycles; unclear root cause.
- Proposal: Cut discovery time by predicting reasons from prior interactions across channels.
- Tech: Page 2 simulates agent desktop with “Reason for the call” recommendations with weights.
- Flow: Interviewer picks, you role-play the agent, start with specific questions instead of “How can I help?”
- Story: POC-ready, multi-source ETL plus AI reasoning, leads into compute sizing and ROI.