Module 5, Lesson 2
This is your full rehearsal. You're going to run through the complete AI-assisted work cycle on a simpler task — the same process you'll use in the capstone, but with lower stakes. By the end, you'll have a "mini Real Tuesday" with all 6 sections completed.
This is the workflow you've been building piece by piece. Now you'll run it end to end.
Step 1: Pick your task. Choose one:
Step 2: Assess the task. Place it on the reliability spectrum. Name: output type, verification method, failure consequence.
Step 3: Classify data sensitivity BEFORE submission. List every piece of information you plan to share with the AI. Classify by sensitivity level (Public, Internal, Confidential, Restricted) and pick the appropriate service tier. If you have Confidential or Restricted data, decide your mitigation NOW: anonymize, omit, or use a paid/enterprise tool. This is a pre-submission gate, not a retrospective check.
Step 4: Select your tool category. General chatbot, research engine, document-grounded, or specialist? Why? Verify your selection works with the data tier from Step 3.
Step 5: Construct your four-element prompt. Role, context, format, constraints — all labeled. Include only the information you cleared in Step 3.
Step 6: Submit and save the complete, unedited output.
Step 7: Diagnose. Find 3+ failures with quoted text, named category, and consequence.
Step 8: Evaluate against professional standards. Name 2+ standards relevant to this task. Find 3+ revisions with original text, revised text, standard cited, and rationale.
Step 9: Revise and produce your final version. Apply all revisions. Save the final artifact alongside the original AI output.
Your mini Real Tuesday should have all of these (same 9-box rubric the Capstone uses):
If you completed all 9 steps without getting stuck, you're ready for the capstone. If you got stuck at any step, go back to the relevant module before attempting the capstone:
Your domain expertise is what makes this workflow work. The AI generates drafts; you know what "good" looks like in your field. That combination — AI speed plus human judgment — is what the capstone proves you can do.
You've completed the full cycle once. The capstone is the same process on a task that matters — something you actually do every week. One last step: choosing that task and setting your baseline.