Building WeMachines on WeMachines
Listen to this article 2 min
We use WeMachines to build WeMachines. This account describes how our four-person team works with AI on tasks, docs and internal tools. The time savings below are our estimates, not measured results for other teams.
Transparency
Our channels, tasks and docs are visible to the team. That lets people and machines follow the same discussions. We keep decisions in Docs so they are available after the chat expires.
Tickets
We ask AI to create tasks from conversations and update them when the plan changes. Machines can use the workspace context we share with them. We still need to give clear instructions and review their work.
We ask the AI to create and update tasks from chat. GitHub workflow rules move them as code is pushed or merged. We check the board when we need to review the work.
Bringing tools together
Before WeMachines, we used separate tools for chat, tasks, docs and calls. Connecting them took time, and someone had to maintain those connections.
Now we use one workspace for those jobs. We also use Apps for internal tools, including a CRM and connections to PostHog, ElevenLabs and our backend. WeMachines handles hosting, and we test and improve the tools in their threads.
AI models
We use different models for different jobs. The model we choose for a quick update may differ from the one we use for a coding task.
WeMachines lets us try models from several providers in the same workspace. Each machine can use its own model, and the team shares one credit balance.
Shared context
When conversations, plans and code sit in separate tools, an agent needs access to the relevant parts. Keeping that context current takes work.
In WeMachines, machines can read our chat, tasks, docs and connected repo. We choose which docs to include. It still helps to point a machine to the task or decision that matters.
Working with coding agents
Our coding agents connect through MCP, so they can use the same tasks and docs as the team. Their current work appears on the board. We save lasting decisions in Docs and keep code review in GitHub.
Our estimated time savings
We estimate that WeMachines saves our four-person team roughly 40 to 50 hours a week across these activities. Your results will depend on your tools and how you use them.
We spend some of that time discussing ideas, reviewing work and talking to each other. Those conversations help us decide what to build next.
What we have learned
We use the product every day and improve it when we find a problem. Working this way has changed how we write tasks, use AI and build internal tools.
Start with one project and see which parts help your team. Keep the decisions clear, review the output and adjust the workflow as you go.
