<?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/4af7911170fb4fa693c37cfc68f69172&quot; frameborder=&quot;0&quot; width=&quot;1112&quot; height=&quot;834&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>834</height><width>1112</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>834</thumbnail_height><thumbnail_width>1112</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/4af7911170fb4fa693c37cfc68f69172-18a165011a718fa6.gif</thumbnail_url><duration>238.233</duration><title>AI Agents for Product Design Validation</title><description>This Loom shows how a product manager can use a multi-agent AI workflow to generate and validate UI design decisions using business analytics without waiting on designers. The creator built four Switchboard agents: Brief to tighten a PRD by asking questions, Designer to produce pixel-perfect Figma-based variations via Figma MCP, Critic to evaluate the designs, and Keeper to log decisions. The process helps teams validate decisions in about half an hour to one hour. A live example demonstrates the PRD tightening, design generation using a pre-made design system, and critique that recommends which design best fits.</description></oembed>