One Agent, Many Sessions

How many AI agents do you actually need, and what changes when other people share the same one?

Hannes Rudolph is joined by Jacqueline Henriksen (Ritz) and Vitor Cepeda Lopes to talk about how they use OpenClaw, both on their own and with other people. They discuss the contributor team’s shared RoboClaw server, where visible sessions and shared context help people coordinate fixes instead of copying messages between agents or duplicating work.

The conversation moves from team collaboration to everyday workflows: email and calendar support, creative work, household administration, and writing an agentic engineering playbook. They compare one main agent with many sessions against separate agents or gateways for privacy and restricted access, and distinguish those choices from handing parallel tasks to subagents.

They also explore when messaging channels are useful, when the web interface offers needed visibility, and what would make OpenClaw more approachable. A recurring theme is verification: clear goals and evidence of useful work matter more than how long an agent can keep running.

Join the OpenClaw Discord. Watch for the next live recording announcement.