What to do on Monday

Do not begin with an organization-wide rollout. Begin with one owner, one task, one evidence pack, and one approval gate. The purpose of Monday is to create a trustworthy learning loop.

1. 09:00 — choose a low-risk recurring job.
2. 09:15 — write its OCEA brief and acceptance test.
3. 09:35 — prepare a minimal approved evidence pack.
4. 10:00 — run at medium effort on the sensible default tier.
5. 10:30 — review facts, sources, structure, and permissions.
6. 11:00 — make one controlled revision and compare.
7. 11:30 — record time, cost, errors, and reusable instructions.
8. Friday — decide whether to stop, repeat, or expand.

If the workflow succeeds, turn the edited brief and review checklist into a template. If it fails, keep the evidence and diagnosis. A documented failure is more valuable than an unreviewed success.

PROCESS IN builds practical systems around real work. Keep the system small enough to understand, strict enough to trust, and useful enough that people choose to use it.

Workshop: What to do on Monday

The practical objective of this chapter is turning one controlled run into a repeatable operating habit. Before opening a chat or Work, write down how the job is performed today, who owns it, and what counts as an acceptable result. Choose one example you can personally inspect. This baseline prevents the novelty of the tool from being mistaken for real improvement and gives you a fair comparison for elapsed time, accuracy, corrections, and usefulness.

Worked example: a department owner spends one morning briefing, running, reviewing, and recording a source-bound weekly update before deciding whether to repeat it. The team first narrows the outcome and assembles this pack: one OCEA brief, a minimal evidence pack, acceptance checks, a correction log, and a Friday decision note. It then runs the agent with an explicit stopping condition, checks every material claim, and records corrections. The example succeeds only when a named owner can explain why the result was accepted, which parts remained human work, and exactly what should change on the next run.

The most common misapplication is announcing a broad transformation before one team has demonstrated a trustworthy review loop. The correction is not automatically a stronger model. First reduce scope, rank sources, reveal hidden constraints, and add an approval checkpoint. If the problem remains after those changes, test a higher tier or effort setting on the same sample. This allows you to compare cause and effect instead of merely comparing two different-looking outputs.

• Describe the current human process in five sentences.
• Prepare an evidence pack that includes one OCEA brief, a minimal evidence pack, acceptance checks, a correction log, and a Friday decision note.
• Set one measurable acceptance rule and one prohibited action.
• Ask the reviewer to classify errors by severity, not only style.
• Save the brief, settings, output, and corrections as one run record.