video1.0<iframe src="https://www.loom.com/embed/10c7a03ef43040eb87ed05268e0d6d56" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/10c7a03ef43040eb87ed05268e0d6d56-d08cce17b0249221.gif3812.96This Loom explains how to craft better generative AI prompts for research administration to reduce exhausting back and forth. Irma Roman and presenter Trisha Hallahan introduce a CRAFT framework: context, role, audience, format, and task, emphasizing an iterative, conversation-based approach rather than a static search. They stress privacy and compliance considerations such as using institutional licenses and verifying data retention settings, and they warn to fact check because AI can hallucinate. Using examples like analyzing funding opportunities and reviewing a notice of award, they show how adding context, specifying the audience, and defining output format improves results. A question-and-answer session also mentions looking for community resources, such as NAURA’s AI collaborative group, though no specific prompt directory was confirmed.