Supervising AI is the lawyer's professional obligation, but supervision is not a single act. It is a workflow. This webinar introduces DIAL, a four-stage supervision framework for AI-augmented legal work: Deliberate, Iterate, Audit, and Log. The Deliberate and Iterate stages are the input-side disciplines, where lawyers shape what they ask AI to do and refine the result. Audit and Log are the output-side disciplines, covered in depth in the AI Navigator Workshop.

This webinar covers Deliberate and Iterate in practical detail. Participants will learn how to construct prompts that supply the context AI cannot infer, including clarity, situational constraints, and role assignment. Confidentiality is treated as a Deliberate-stage discipline because once information enters a prompt, the disclosure has already occurred. The session then walks through five iteration techniques that work across any AI platform and any legal task: one-shot prompting, few-shot prompting, chain of thought, tree of thought, and counter-prompting. A specific legal prompt architecture grounded in rhetorical situation theory will be introduced and explored in the AI Navigator Workshop.

Learning outcomes. 

After this session, participants will be able to:

1. Describe the four stages of the DIAL supervision framework and identify which stages this webinar covers;
2. Construct prompts that supply the clarity, context, constraints, and role assignment AI requires;
3. Apply five iteration techniques (one-shot, few-shot, chain of thought, tree of thought, and counter-prompting) to refine AI output; and
4. Perform an input-stage confidentiality check before sending any prompt, consistent with Louisiana Rules of Professional Conduct 1.6, 5.1, and 5.3.

This is the third and final webinar in the AI Explorer Series, which establishes the foundation for responsible AI use in legal practice and is required for participants planning to attend the AI Navigator Workshop.