During a drive yesterday, my son complained "Why is the road so bumpy? I'm not able to draw!", on a small patch between highways.
I mentally facepalmed. Growing up in a tier-3 town in India, every road was bumpy. We couldn't 'draw' or 'read' on drives. We were 'forced' to look outside and notice everything.
When you have a 12-lane smooth highway in a big city, the experience is very different from a bumpy 4-lane street in a town. There are no shops to see next to the road, no impulse stops for a forgotten errand. The destination is set, the route is set, the map directs any deviations.
One of my favourite CS profs made this point once, when discussing user journey design: Sometimes, friction needs to be inserted into a user journey, so that users can have a more enriching experience.
For human users, having some friction in how a UX is designed is desirable, to allow for a feeling of being in control. When designing the same UX (or AX) for autonomous agents, these frictions can have severe negative consequences, with the user agent getting confused due to the 'bumps' on the road and forgetting where they were driving. AX, or Agent Experience, cannot be treated the same as UX.
Looking at one of my recent personal projects, I don't see any evidence of Claude Code having a friction philosophy at all, intentional or otherwise. I suspect the coding agents today don't have any guidance on how to handle friction. I also suspect most users of these coding agents aren't giving this guidance either (I know I didn't, for this project).
Two questions here:
- If you're designing human-user interfaces using AI-generated code, who is responsible for the friction philosophy?
- If you're building for autonomous agents, are you bringing over your UX assumptions (including added friction), or thinking about AX as a distinct problem?
#UXDesign #AgentExperience #AIAgents #ProductDesign