2.1.2. Interaction Surfaces: Inline, Chat, CLI, and Plan Mode
💡 First Principle: Each interaction surface is a different bandwidth for talking to Copilot. Inline suggestions are the lowest-bandwidth, fastest channel; Chat is conversational; the CLI brings Copilot to the terminal; Plan Mode adds a deliberate "think before acting" step for complex work.
Within the IDE you can trigger Copilot several ways, and the exam expects you to match the trigger to the need:
- Inline suggestions — ghost text completing the current line or block as you type. Accept with Tab, reject by continuing, or cycle alternatives. Best for staying in flow while writing known code.
- Copilot Chat — a conversational panel for questions, explanations, test generation, and debugging. Best when you need reasoning or back-and-forth, not just a completion.
- CLI triggers — invoking Copilot from the terminal (covered in depth in 2.2). Best when your work is command- or script-shaped.
- Plan Mode — Copilot drafts a step-by-step plan before making changes, so you can approve or adjust the approach. Best for multi-step tasks where you want to see the strategy before code changes land.
The progression mirrors the autonomy ladder: inline for a line, Chat for a question, Plan Mode for a strategy you'll then execute (often in Agent Mode).
✅ Best Practice: Reach for Plan Mode when a task is big enough that a wrong approach would be expensive to unwind. Seeing the plan first is cheaper than reviewing a sprawling set of edits after the fact.
⚠️ Exam Trap: Plan Mode does not itself complete the work — it produces the plan. Execution still happens (commonly through Agent Mode). A question that treats Plan Mode as the thing that writes all the code is describing it incorrectly.
Reflection Question: A developer faces a multi-step refactor and wants to review Copilot's intended approach before any files change. Which surface fits, and what does it produce versus an execution mode?
