You are the engineering manager leading a project to redesign a customer support workflow for a B2B SaaS analytics platform that handles advertiser and partner issues across attribution, SDK setup, fraud alerts, and billing. Support volume has grown 40% in two quarters, first-response SLA is slipping, and leadership wants more automation to reduce repetitive work, but several high-value enterprise accounts have recently escalated because automated responses missed account-specific context. You need to decide what should be automated versus kept manual in the first release, while coordinating support operations, backend engineering, data, and security. The project is tricky because the support team wants immediate relief, the enterprise success team is worried about customer trust, and one of the key dependencies is a case-routing rules engine that is already shared with another roadmap commitment.
| Detail | Value |
|---|---|
| Support tickets per month | 18,000 |
| Tickets in scope | 4 categories covering 72% of volume |
| Engineering team | 4 backend, 2 frontend, 1 data engineer |
| Support operations team | 1 manager, 6 agents for pilot region |
| Deadline | 10 weeks to pilot |
| Budget | No additional headcount; $60K vendor/tooling cap |
| SLA target | First response under 2 hours for priority tickets |
| Compliance constraint | No automated outbound response may expose customer-level attribution data without approval |
How would you decide what to automate versus keep manual, and how would you plan and execute the pilot so you can improve support efficiency without increasing customer risk or harming SLA performance?