Case 01
Stabilising scope on a live decision-support platform
An enterprise decision-support platform was already in delivery and its scope kept growing.
- Where
- AI-CoE for Sustainable Cities, India
- When
- 2025 to 2026
Context
An enterprise decision-support platform was already in delivery when I joined, and its scope kept growing with every meeting.
My task
Write down what the platform had to do, and get everyone to agree what ships first.
How I did it
I wrote the business and functional requirements from the workflows people actually followed, not the ones in the slides. Every requirement got an ID, a source and an owner, and open questions stayed visible on a list instead of being guessed. Then I ran a MoSCoW prioritisation with the team, so 'must have' meant the release fails without it.
Research and insights
- 01
Scope creep is usually unanswered questions in disguise. Writing them down, with an owner each, stopped them coming back as new features.
- 02
Mark every rule as confirmed or proposed. A suggestion that looks like an agreement is how a spec quietly grows.
- 03
Trace every screen to a requirement ID. If a screen has no requirement, either the spec or the screen is wrong.
How it was delivered
A requirements set the team could build and test against, and an agreed first release. Scope stabilised.
Step by stepEvery step, with what each one produced
- Join the live delivery
Came in while the platform was already being built.
- Write it down
Business and functional requirements for the platform.
- Agree what ships first
MoSCoW prioritisation with the team.











