Expensive Certainty: The Step Before It Becomes Code

Draft a clear statement of what your unit produces, for whom, and to what standard. Test whether the work matches it.

Trace it to the organizational mission. Where the line breaks is a finding that usually appears before anyone examines the tool.

Watch it happen first in the simulation.

Separate the obvious answer from the real constraint while the simulation lets the wrong diagnosis play out.

Watch what the wrong diagnosis costs when the situation answers back.

Apply the same distinction to a live problem in your own unit between sessions and bring back what it surfaces.

The knowledge is not missing. It sits with the people doing the work. Nobody has ever asked them to state it.

Map the people who know what actually happens across levels and functions. Not the org chart. The real sources.

Practice the question that produces a usable answer instead of a plausible one.

Surface the overrides, work-arounds, and unwritten calls that shape the work.

Ask what each one is protecting. The original reason may still be valid even if no one wrote it down.

Sort them: real protections, old habits, and decisions no one can explain anymore.

Hold each surfaced decision against the purpose statement from week one. Does it serve what the unit exists to produce, or something else.

Decide which deserve to survive. Name what is lost if you retire it and what is risked if you keep it.

Identify which decisions sit above your level and cannot be resolved inside the unit.

Document what is broken, what it costs to leave alone, and which decisions sit above your level.

Use AI to check the document for places where an opinion was stated as a finding.

Leave with the method, not only the document. You will run it again the next time a decision arrives without you.