Demonstrate and hand over

Outcome and preparation

Day 2 · 16:15–17:00 · 45 minutes. Demonstrate the MVP and hand it to a colleague. Bring the handover folder and local routine history. Allow 5 minutes to prepare, 30 for short demos/peer checks and 10 for commitments. With 10–12 learners, keep each demo near two minutes and use paired checks for details.

Completion means you can show the work and its limitations. A polished report, saved prompt or configured schedule is not sufficient evidence that the whole workflow operates.

Step by step

1. Step 1

   Inventory the handover folder: task/approach cards, prompts, Project instructions and refresh checklist, verified report, Skill, agent canvas/file/tests, routine configuration/run evidence, business plan, ROI scenarios, three pilots and operating policy. Use the prompt to identify gaps.

   Prompt / exampleCopy

   Review this handover inventory against our executive-reporting MVP acceptance criteria.
   Separate verified outputs, configured-but-not-tested items and missing evidence.
   Do not mark an item complete merely because a file or schedule exists.
   Produce a concise checklist with owner and next action for every gap.
   [Paste inventory and rehearsal log]
   

2. Step 2

   Give a two-minute demonstration: show the original decision and inputs; run the Skill; open the evidence and report; verify one calculation and one source locator. Show the unchanged rerun. Do not present a prebuilt answer as a fresh agent run.

3. Step 3

   Show a failure case: stale source, missing file, conflicting revenue or malicious source instruction. Explain where it stops, what the user sees and who resolves it. State which tests were actually run on the agent versus the deterministic calculator.

4. Step 4

   Open routine history. Show the actual scheduled session timestamp and output status, then the current Paused state. If no clock-triggered run occurred, label this item pending and book the remaining rehearsal rather than claiming completion.

5. Step 5

   Present the business plan's decision, budget assumptions and three pilot commitments. Report before/after preparation AND review time. Distinguish observed cost, allocated subscription cost and unknown usage. Name one piece of evidence that would make you stop.

6. Step 6

   Ask a partner to follow your handover without verbal help. They should locate the sources, run the historical fixture, inspect output and explain pause/recovery. Record misunderstandings, fix the instructions and save handover/final-review.md with reviewer, date, passed checks and open items.

7. Step 7

   Write a Monday action: who will do what, with which approved data, by when. Confirm the practice routine is paused. Keep real-source connection and production adoption behind the pilot's review gates; do not leave a fictional weekly job running indefinitely.

Expected output and verification

An inspectable handover pack and a clear status for every acceptance item. The workshop records participation and learner-marked progress; it does not confer accreditation or automatically certify production readiness.

Minimum demonstration: correct report period and calculations, evidence citations, explicit uncertainty, real Skill/agent invocation, repeat-run handling, an observed local scheduled execution, human approval boundary and a costed pilot plan. Any missing item remains a visible follow-up.

Troubleshooting

If time runs out, save evidence and have your partner finish the review after class. If a live run is blocked, show a labelled backup while retaining the blocked status. If another person cannot reproduce your steps, fix paths, prerequisites or missing permissions before describing the workflow as handed over.

Completion checklist

• Peer can run and verify the workflow from your notes.
• Normal, failure and scheduled execution evidence distinguished.
• Business plan, three pilots and Monday action reviewed.
• Practice routine paused; unresolved items assigned.

Return to the workshop overview for the source mapping, prework and downloadable pack. Reuse the method on approved organisational data only after your pilot's access and review arrangements are in place.