Project Context
Northstar Pay, a mid-market payments platform, is seeing a sharp increase in refund requests after launching instant merchant onboarding. The current refund review process is fully manual: 18 operations analysts review roughly 4,500 requests per week, and average decision time has risen to 31 hours. Leadership wants to reduce backlog before the holiday peak in 10 weeks.
A cross-functional team of 9 people has been assigned: 4 engineers, 1 data scientist, 1 designer, 1 operations manager, 1 compliance lead, and you as program manager. The core decision is whether to invest in workflow automation now, keep manual review with temporary staffing, or launch a hybrid model.
Key Stakeholders
The VP of Operations wants immediate backlog reduction and is pushing for temporary contractors. The Head of Engineering prefers building automation to avoid scaling headcount. Compliance requires that any automated decision path remain auditable and keep false approvals below a strict threshold. Finance has approved limited budget and will not fund both a full automation build and a large contractor ramp.
Constraints
- Deadline: 10 weeks before holiday transaction volume increases by 40%
- Budget: $220,000 total
- Engineering capacity: 4 engineers, with 1 shared at 50% due to another launch
- Current SLA target: 95% of refund requests decided within 12 hours
- Automation dependency: model-based risk scoring service is available, but only has 82% precision in pilot results
- Contractor onboarding time: 2 weeks minimum
Complications
- A recent audit found inconsistent documentation in 12% of manual refund decisions.
- Sales has promised two enterprise merchants that refund turnaround will improve by next quarter.
- The risk scoring service owner says precision may improve, but retraining will take at least 4 weeks.
Deliverables
- Recommend a manual, automated, or hybrid execution strategy and justify the trade-offs.
- Build a 10-week execution plan with milestones, owners, and launch criteria.
- Define how you would decide which refund cases stay manual versus move to automation.
- Identify the top risks, including compliance and delivery risks, and propose mitigations.
- Specify success metrics for backlog reduction, SLA performance, and decision quality.
