Designing Better Human–Agent Collaboration
You already have agents. They can research, write, code, analyze, design, and execute. As agents become more capable, the challenge shifts towards figuring out how humans and agents can effectively collaborate
How do you actually work well with agents? Where does collaboration create the most leverage? What should we delegate? How much context does an agent need? What changes when multiple agents are working alongside a team? How do we review its work without becoming the bottleneck?
In this hands-on workshop, Pablo Stanley, Staff Product Designer, Vercel & Founder of Hilos will explore the emerging patterns for effective human–agent collaboration and apply them to the agents you already use.
___
During the workshop, we’ll work through practical ways on how to collaborate better with agents as well as try out Hilos a collaboration chat where humans and AI coding agents work in the same rooms. We'll
Delegate outcomes — how to give agents clear goals, context, constraints, and definitions of done.
Design effective collaboration loops — how to move beyond prompt → response toward delegate → act → review → iterate → approve.
Build the right level of autonomy — how to decide when agents should act independently, keep you informed, or stop for approval.
Coordinate multiple agents — explore delegation, specialization, shared context, handoffs, and escalation across human–agent teams.
___
About Pablo Stanley
Pablo Stanley is a staff product designer at Vercel, working on v0. He's spent his career building tools for creatives: Blush, an illustration tool, Lummi, a free AI stock photo library, and design work at InVision Studio. He writes The Design Team webcomic and a newsletter about AI, design, and creative work. Right now he's building Hilos, a chat where humans and AI coding agents work in the same rooms.