When you have a tool that can reason, analyze, synthesize, generate ideas, and automate tasks, it’s tempting to point it at a problem and start redesigning things.

I see this happening with work. There’s a growing conversation about work architecture, breaking jobs into tasks, reorganizing workflows, and rethinking what people and machines should do.

As an architect (buildings!), I’m intrigued, and it brings to mind the kinds of questions you might hear during a desk crit in an architecture studio.

• What’s your big idea?

• What are you actually trying to achieve?

• What’s the organizing principle behind your design?

• Why this approach and not another?

• What assumptions are you making about how people behave?

• What does a person experience moment to moment, day to day?

• Where do people encounter one another, and what happens in those encounters?

• What relationships does your design create, strengthen, or eliminate?

• What are the key forces shaping your concept?

• What happens at the edges, where one activity meets another?

• Where does the design allow for adaptation, improvisation, and the unexpected?

• What happens when people use it differently than you intended?

• What are you optimizing for, and what might you lose in the process?

• How will you know whether the design actually works?

• And perhaps most importantly, what kind of human experience does your architecture make possible?

These aren’t questions about technology. They’re questions about design.

AI gives us extraordinary new capabilities to rearrange the components of work, but rearranging something isn’t the same as knowing what it should become.