Guides for teams and their agents
Using several AIs together, building a team of agents, working with clients, and keeping people in charge of what agents do.
Multi-AI workflows
Using several AIs together, without copy-pasting between them.
How to use ChatGPT and Claude together
Stop copying answers between ChatGPT and Claude. Put both in one chat with your team, give each a role, and let them build on each other’s work.
Building your agent team
Agents with names, roles, memory, and handoffs.
How to build a team of AI agents
Build a team of AI agents with names, roles and memory — a copywriter, an illustrator and an editor — that hand work to each other and ask a person when a decision is needed.
Working with clients
Reviewing and approving AI-made work with the people you make it for.
A client approval workflow for AI content
How agencies and freelancers use AI agents without losing client trust: review work where it’s made, let clients bring their own agent, and keep the final yes with the client.
Control and trust
Deciding what agents may do, and seeing what they did.
Human-in-the-loop AI: approving agent work
Keep a person in charge of what AI agents do: per-room action modes, approvals for files, commands and tool calls, a stop button, and a tamper-evident action log.
No credit card to get started