TL;DR
A centralized Indigenous Consultation repository served 31 Ontario ministries, and enhancement requests arrived from many of them at once. My division had a short, fixed window to deliver. I decided what made the cut, then ran the delivery end to end. Consultation processing time came down 20% against a three-to-six-month baseline.
The problem
Thirty-one ministries, a contracted engagement with a hard end date, and more requests than the window could hold. Every request was legitimate to the ministry that raised it.
What I owned
What made the cut, and the delivery after it. Agile delivery end to end: user stories and acceptance criteria through to functional, regression, integration and user acceptance testing. Government delivery has a deserved reputation for slipping, and the testing depth is a large part of why this one did not.
The hard part
Deciding what to leave out. What actually drove it was effort against time remaining. Not a north-star metric, not a strategic ranking. In a fixed-window contracted delivery, capacity is the binding constraint, and pretending otherwise produces a beautiful prioritization framework and a missed deadline. Five features fit, so five features were what got committed. The skill on display is triage under a hard limit, which is a real and common situation, and I would rather describe it accurately than dress it up.
Outcome
The five-feature scope delivered inside the window, and consultation processing time cut 20% against a baseline that had been running three to six months. Across 31 ministries, that compounds.