• Audience: Beginners familiar with submitting prompts; no technical background required.

• Goal: Produce a stronger, fact-preserving draft using a shared rubric, separate AI reviews, targeted revisions, and a defensible human acceptance decision.

• Narrative mode: Tutorial.

• Slide count range: 15 slides proposed.

Suggested session: 35–45 minutes including practice; instructional estimate, not a verified five-minute method.
Running example: A fictional operations manager announcing a move into workflow consulting on LinkedIn.
Companion handout: Detailed tutorial and teaching guide. Full prompts, scoring anchors, checkpoints, and implementation checklist stay in the handout.
Source: My Simple But (WEIRD) Way of Getting Better ChatGPT Results, full available transcript, 0:00–8:38. Visuals were not independently inspected. All exercises and operational rubric definitions are course additions.

1) Improve the draft—not just its score

Turn one career-change announcement into a draft you can defend with evidence. Promise a repeatable review method, not guaranteed results or a perfect score. Introduce the presenter's “force prompting” as the source concept; teach explicit criteria and human oversight. Finish with a baseline, reviewed revision, and a brief acceptance rationale.

2) A confident self-rating can hide generic writing

• Contrast the presenter's reported 9/10 self-rating with his objections to “safety net” and “tectonic shift.”

• Ask what those phrases actually tell the reader about this person's decision.

• Separate an enthusiastic rating from evidence that the brief was met.

• Anchor to 1:05–2:35; frame the example as presenter-reported, not a model benchmark.

3) Follow the learning path from brief to decision

1. Define the audience and preserve a baseline.

2. Set standards and obtain separate, evidence-based reviews.

3. Revise, investigate disagreement, and compare versions.

4. Accept the draft only when a human can justify the result.