Building AI Agent Workflows with Switch

​Bring existing AI agents into the places teams already work, with shared context and history - Louis Amaudruz, AI Engineer at SandboxAQ

​In this hands-on workshop, we’ll build a multi-agent workflow with Switch, an open-source workspace where people and AI agents can coordinate real work through governed rooms. We’ll start with a real use case, connect multiple agents, and work through the solution live to see how agents can coordinate with people while maintaining shared context.

​We’ll cover the following steps:

• ​Define a real-world workflow that benefits from people and AI agents working together.

• ​Build the workflow in Switch, bringing existing agents into a shared room and putting them to work alongside people.

• ​Keep context with the work, exploring how rooms preserve context and history as people and agents join, leave, and hand off work.

​By the end of the workshop, you’ll be able to build your own AI agent workflow in Switch, bring existing agents into a shared room, and use persistent context and history to keep people and agents working together.

​About the speaker:

​Louis is an AI engineer at SandboxAQ, the parent company behind Switch. He and his team built Switch to solve a problem they experienced firsthand: as they adopted more AI agents across different tools and frameworks, putting those agents to work across a team required new integrations and repeated context sharing. Switch grew out of that need, bringing people and agents together within the collaboration tools teams already use while preserving context and history of the work.

​DataTalks.Club is the place to talk about data. Join our Slack community!