Positioning you as the supervisor and final editor, this session explains how generative models actually “draft” so you can manage them like a perpetually junior associate. We’ll distinguish linguistically correct from factually correct, show why hallucinations occur, and flag Louisiana-specific pitfalls (civilian terminology, parish procedures, statute formats). 

Framed as managing a junior drafter, we’ll explore common prompting tactics—role prompting, few-shot examples, tone/length controls, and step-by-step instructions—as “simple assignments” you give the AI. You’ll learn where these strategies fit in the workflow (brainstorming, outlining, first-pass drafts) and how to supervise them to avoid scope drift and jurisdictional errors. You’ll leave knowing when basic strategies are sufficient and when your supervisory role requires escalating to a structured approach, tighter guardrails, or legal-vertical tools.

Advancing the supervisor–editor model, this session introduces GATEPAS as your “design brief” for delegating legal drafting to your digital assistant. We’ll walk each element—Genre, Author, Tone, Event, Purpose, Audience, Subject—showing how it clarifies the assignment, sets performance metrics, and anchors jurisdiction and audience. You’ll see before/after outputs and learn to encode success metrics, assumptions, and uncertainty so your editorial review becomes faster and more reliable.

You’ll practice delegating like a supervising attorney: craft a complete GATEPAS “assignment” for an anonymized matter, specify the audience and jurisdiction, and include guardrails the AI must follow (clarifying questions, scope limits, uncertainty statements). Using a fillable cheat sheet and peer checklist, you’ll produce and save a versioned v1.0 prompt to your shared library—ready for you, the editor-in-chief, to test and refine.

Here you perform the editorial pass and quality control. Run your v1.0 prompt, then use the Accuracy–Bias–Confidentiality–Judgment checklist to mark what the AI got wrong or missed. Add verification requirements, tightened scope, and document assumptions. Swap prompts to check reproducibility—an essential supervisory test—and update notes to v1.1. The result is an assignment-and-review loop you can repeat.

Turn your supervisory workflow into organization infrastructure. We’ll set up a prompt library in Notion or Excel/Google Sheets. You’ll leave with a seeded library, a governance plan, and a 30-day roadmap to institutionalize the supervise–edit–verify workflow across your team.