Turning a plan into working documents

A compelling plan is not automatically an executable one. It needs to connect how a place will operate, which decisions come first, and who owns them before it can guide day-to-day work.

UPMI's public urban-regeneration report addresses the transition from selection to delivery. The Lab uses this as a document-structuring problem, distinguishing preparatory tasks suitable for agents from decisions that require people.

The first record is an action plan with dependencies. The second maps responsibilities and reviewers. The third records stakeholder input and why it was accepted or declined. The fourth links claims to source material.

Using consistent task names and reference dates makes the records easier to compare. An agent-generated table should also distinguish missing owners and unconfirmed dates from established facts.

The published report proposes a 30-, 60-, and 90-day framework that progresses from initial issues and responsibilities to operating arrangements and delivery controls. These periods are a planning aid, not a mandatory schedule or a claim of completed UPMI work.

A real project needs a schedule based on its consultation and approval requirements. A date on a chart is not evidence of readiness; the relevant question is whether the next task has what it needs to begin.

Start with a short plan that can be shared. Ask an agent to extract tasks, owners, timing, and supporting evidence into a table, leaving missing values explicitly unconfirmed rather than filling them by inference.

A human comparison against the original can then identify omissions and misunderstandings. This creates both a deliverable and a reviewable work record. The page proposes an experiment; it does not promise performance on a particular project.