About
Alfred Poindexter
Principal UX and product designer. I build and manage teams of AI agents, and I write about what it takes to run them well.
My background is product strategy, user research, interaction design, experimentation, and design systems. Most of my career has been spent turning messy customer and business needs into requirements, working prototypes, and acceptance criteria that engineering teams can ship against.
Today I apply that same discipline to AI-native products. I design and build multi-agent systems where specialist agents each own one narrow job, a coordinator moves work between them, and the decisions that matter stay with a person. I’ve also built an AI mystery-shopper audit system that keeps the agent doing the work separate from the code that checks it.
The question I keep coming back to: how do you give agents enough room to be useful without giving up the judgment that makes a product trustworthy? That question drives my interest in human-agent alignment, evaluation, and governance, and it’s what this writing is about.
What I write about
- Division of labor. Why bounded specialist roles beat one generalist agent in production.
- Orchestration. How work moves between agents, and why the coordinator should be the last role to earn autonomy.
- Human governance. Designing checkpoints, stops, and review screens so that oversight is real, not a rubber stamp.
The best place to reach me is LinkedIn.
Get the next essay when it ships.
One piece at a time on building, managing, and governing AI agent teams. No hype cycle.