Presentation

The final skill this course certifies isn't prompting — it's being able to brief others credibly. The person who can explain what AI did, what it didn't do, and where the human stays in the loop becomes the team's trusted voice on the subject. That's a career asset. Your briefing has five beats.

1. The task and its old cost (30s). 'Every Monday I wrote the ops summary. It took ninety minutes.' Concrete, relatable, no AI vocabulary yet.
2. The workflow in one picture (60s). Show the station diagram. Emphasize the human checkpoints — this is what earns trust from skeptics.
3. A real before/after (90s). One actual input, the AI's raw extraction, your correction at the checkpoint, the final output. Showing the correction is the masterstroke: it proves the safety story instead of asserting it.
4. The honest numbers (60s). Time saved including review; what the checkpoint has caught; the failure modes you know about ('it misses context from calls, so call notes get added by hand').
5. What it would take to adopt (60s). The one-pager exists; the prompts are copy-paste; here's the data rule to respect. End with the offer, not a sales pitch.

Build it fast, own it fully

Handling the two questions you'll definitely get

• 'How do you know it's not making things up?' Don't say 'it's usually right.' Show the checkpoint: 'I review the extracted facts before they're formatted — here's one I corrected, and here's my spot-audit cadence.' Process, not faith.
• 'Could it do my [other task]?' Resist instant yes. Run the four filters from the brief out loud — recurring? 30+ minutes? green-zone core? your data class? You'll either design their workflow on the spot or explain precisely why not. Both answers build your credibility.