From a problem to a useful next step.

First, understand the work. A description of a recent task, what made it difficult and who felt the effect is often more helpful than a proposed solution alone.

Then, define a bounded outcome. Discuss what would count as useful, the constraints that matter, and the smallest step that could test the idea. Discovery may show that an existing tool is sufficient.

Before paid work begins, agree the scope, responsibilities, budget and delivery arrangement. This page describes an intended approach, not a fixed service package or a promise of availability.

Review what was built against the real flow, including failures and recovery. Keep what helps, correct what does not, and decide the next step using that evidence.