Why does a multi-agent design need a clear coordinator or handoff?
Why B is the answer
So that tasks and context move to the right agent
Clear handoffs stop work being dropped or duplicated.
In detail
- A clear coordinator moves tasks and context to the right agent.
- Say which agent owns each step.
Use case
An illustrative scenario, not a real incident.
Two agents each think that the other is handling a refund, so nothing happens. A clear coordinator prevents that.
See it in different ways
flowchart TD A["User"] --> B["Coordinator agent"] B --> C["Sales agent"] B --> D["Finance agent"] B --> E["Service agent"]
sequenceDiagram participant U as User participant C as Coordinator participant S as Sales agent participant F as Finance agent U->>C: Request C->>S: Task and context S-->>C: Result C->>F: Task and context F-->>U: Completed work
sequenceDiagram participant P0 as Coordinator participant P1 as Agent A participant P2 as Agent B P0->>P1: Task and context P1-->>P0: Result P0->>P2: Next task and context
Key point: Clear handoffs stop lost work.
Why the other options are wrong
- A. So that users have to sign in several times
- Multiple sign-ins would be a defect.
- C. So that each agent can use a larger model
- Model size is a separate choice.
- D. So that the number of agents stays exactly even
- An even number is meaningless.