A business analyst uses Copilot Chat in VS Code to research a billing bug, draft unit tests, and explore a new checkout flow. She keeps getting generic answers when she needs codebase edits, test scaffolding, and exploratory guidance. Which approach to selecting an agent in chat best matches each goal?

The correct approach aligns each Copilot Chat mode with its designed scope: Ask provides codebase-aware Q&A and explanations without modifying files, which fits researching the billing bug history and behavior, Edit applies focused inline changes to selected files, which fits drafting unit test scaffolding, and Agent mode autonomously plans and executes multi-step tasks across files and terminal, which fits exploratory checkout work requiring guidance plus edits. One plausible but incorrect pattern keeps exploration in Ask to stay safe; while cautious, Ask remains read-only and cannot scaffold or iterate on code, so the analyst would continue receiving generic answers instead of actionable exploration. Another plausible pattern limits Agent to read-only summaries; this misunderstands autonomy, because Agent is specifically built to create, modify, and validate files across the workspace, so restricting it discards the exploratory editing the scenario demands. A third plausible pattern assigns guided multi-file exploration to Edit; this overextends Edit, which excels at targeted changes in context but lacks autonomous planning, tool invocation, and cross-file orchestration needed for open-ended flows. For exam purposes, remember that the distinction hinges on edit scope and autonomy rather than topic, and a practical check is to glance at the mode selector and intended file impact before sending: if no files should change use Ask, if specific files should change use Edit, and if discovery plus changes across files is needed use Agent.