<?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/e15782650d894915a92f3d274605f457&quot; frameborder=&quot;0&quot; width=&quot;1484&quot; height=&quot;1113&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1113</height><width>1484</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1113</thumbnail_height><thumbnail_width>1484</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/e15782650d894915a92f3d274605f457-6fc1f13a5bd81e05.gif</thumbnail_url><duration>532.563</duration><title>How to Use Teams of Agents in Sapience</title><description>This Loom explains how to use teams of agents in Sapience to coordinate multi-agent work. It notes that even a basic chat can automatically invoke teams when a task requires tools such as the Sapience web research agent, which currently uses Google’s AI stack and can alternatively use Perplexity. It then demonstrates the dedicated team-of-agents skill in plan mode, where a 15-minute run used agents to read the website, have the CMO brainstorm messaging, and use the marketing agent to build and publish a new website, with a coordination file as a scratch pad. The video also covers superagents such as CFO and CMO, and a team chat mode where you can create group chats with only agents, switching off the immediate assumption that agents will reply.</description></oembed>