<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/2efc15da468e4ab9a8c49c3af63a8ba0&quot; frameborder=&quot;0&quot; width=&quot;1280&quot; height=&quot;960&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>960</height><width>1280</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>960</thumbnail_height><thumbnail_width>1280</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/2efc15da468e4ab9a8c49c3af63a8ba0-c98dbb8c29e341dc.gif</thumbnail_url><duration>3707.7</duration><title>Understanding AI Agents for Research Administration</title><description>This Loom explains how AI agents can support research administration tasks and why human review remains essential. The presenters introduce Streamline, define AI agents versus chatbots, and cover key concepts like context windows, token limits, and hallucinations, emphasizing clear, contextual prompting and trust but verify. They demonstrate live use cases using tools such as Microsoft Copilot, ChatGPT, Claude, and Gemini, including generating NIH-style budget justification narratives from an R&amp;R budget, summarizing a 50-page FOA with a compliance checklist, drafting sponsor inquiry responses related to SAM.gov excluded individuals, and redlining risky sub-award terms. The session was recorded and includes a prompt template link via a QR code in the recording.</description></oembed>