Tool calls in the UI
Comes up in: Frontend system design, AI system design
This concept is being written. It gets the full five steps — explanation, predict, interview angle, practice with tests, and review cards. Here is exactly what it will cover, so you can study it elsewhere in the meantime.
What it covers
- tool-call lifecycle (requested → input streaming → running → result/error)
- showing tool activity
- approval UIs for risky actions (human-in-the-loop)
- cancelling a tool run
- displaying partial JSON safely
Words you will know
What you will practise
- UI for a "book a meeting" tool with approval
Interview questions this answers
- How would you show tool calls to users while an agent works?
- When should an AI action require user approval? How do you design that UI?