Challenge
Can an autonomous AI agent become a practical teammate within an everyday software development workflow?
AI agents are becoming increasingly capable, but integrating them into real engineering workflows still requires significant configuration, context management, and manual interaction. I wanted to explore what happens when an AI agent participates in the same project conversations, task management, and development processes as a human teammate. More importantly, I wanted to test whether natural conversation could replace some of the repetitive work involved in managing these systems.
Solution
Connecting OpenClaw, Reachy Mini, and Slack to create a voice-driven AI development teammate.
I built an experimental system combining OpenClaw's agent capabilities, Slack's project workflows, and a Reachy Mini robot named Rexi. Reachy Mini serves as the conversational interface, helping translate spoken ideas into actionable requests for OpenClaw. Slack maintains project context, discussions, and task history, while custom skills support development activities. I also created setup scripts, configuration guides, and controls for disabling the robot's audio and camera inputs.
Outcome
A working prototype demonstrates how conversational AI can participate in real software development tasks.
During the recorded demonstration, I used Rexi to resume an existing application project, start its development server, request light and dark theme support, review the proposed implementation, and initiate the changes without typing commands. I verified that the updated interface responded to system appearance settings. The experiment demonstrates a promising workflow, while broader autonomy, repeatable deployment, and integration with my open-source projects remain areas for further testing.