Download the complete part-10 practice pack (ZIP). Extract it and preserve its folders. All organizations, records, and results in this teaching pack are fictional.
These resources are drafts and constructed exercises. They do not approve a tool, adopt a policy, establish savings, demonstrate real recovery, or schedule a rollout.
1. Make one policy decision usable
Open practice/AI-Policy-Draft.md. Complete the organization, owners, approval route, effective date, review date, retention decisions, and approved-register location before adoption. Use practice/Use-Register.csv to specify a workflow's purpose, sources, workspace, and action boundary. Its proposed Mesa example is not approved.
Use practice/Vendor-Evidence.csv to separate a vendor claim from supporting evidence and a business decision. Vendor-A is a fictional placeholder. No contract or vendor document is supplied, and no essential requirement is passed. Replace constructed entries with applicable, dated evidence before making a real decision.
2. Explain one impact scorecard
The Part 8 adoption log and assumptions are copied unchanged into practice/. Read practice/impact-assumptions.json for the additional teaching assumptions: USD 30 per internal hour and USD 6 allocated software cost for the same exercise period.
practice/Impact-Scorecard.csv shows twelve assisted outputs, one hundred total minutes, eight minutes of modeled capacity, USD 56 specified recurring resource cost, USD 4.67 per finally accepted assisted output, and negative USD 2 provisional recurring resource benefit. The modeled cash effect is negative USD 6 under the stated no-cash-benefit assumption. Setup and initial training are excluded; this is not complete ROI or observed business performance.
Use practice/Decision-Gates.csv to document essential conditions alongside economic tradeoffs. Real owners, adopted numerical thresholds, observed results, and approvals remain absent. Do not use the synthetic scorecard to pass a real pilot.
3. Practice continuity without inventing recovery evidence
Read practice/Continuity-Exercise.md, then reveal the events in practice/continuity-scenario.json in order. Use practice/Recovery-Actions.csv to track the post-checkpoint request, uncertain send, and manual intake item. All remain unresolved in this exercise.
The scripted twenty-minute minimum-service interval falls within the proposed sixty-minute RTO on paper. The fifteen-minute export age does not prove the RPO can be met. Application availability at fifty minutes is not full recovery. Record actual technical restoration results separately when a real test is authorized and performed.
4. Plan ninety days with conditions
Adapt practice/90-Day-Plan.csv and practice/Capstone-Checklist.md. Proposed day one is September 14, 2026; day ninety is December 12. All milestones are not started, actual completion dates are blank, and no meetings or automations are scheduled.
A milestone's arrival does not satisfy its prerequisites. Keep scope unchanged until the required evidence and decision are recorded. Choose expand, continue, improve, or stop based on the results.
5. Check the teaching arithmetic
From the Part 10 directory, run:
python3 practice/verify_examples.py
The standard-library checker recalculates the scorecard, verifies inherited fixture bytes, checks the scripted timeline and inclusive dates, and confirms that approval and observed-result fields remain empty. It writes Example-Verification.json. It calls no AI model or external service and executes no business workflow.
For the complete reading sequence, open Core-Guide-Index.md. For the unfinished website audit, open Continuation-Record.md. The forty core manuscripts are complete; the original URL inventory still needs its supporting-page and migration decisions.