GitHub Copilot app: write your first prompt without the perfect phrasing

GitHub has published guidance for developers beginning their first task with the GitHub Copilot app, emphasizing that users do not need to write a perfect prompt to get started. The article walks through the key steps: first, users must give Copilot context by connecting to a project—either a GitHub repository or a local folder on their computer—so the AI can examine the codebase and relevant files. From there, users describe what they want to accomplish in plain English; the example given is "Add a most-funded sort option to the games list," which is already enough to get started. No special syntax is required. If the first attempt does not match expectations, users can provide additional details or ask for changes, treating prompting as an iterative process. The GitHub Copilot app also allows users to choose which AI model handles their task, acknowledging that different models have different strengths—some excel at complex reasoning, others handle simpler tasks more quickly. Users starting out can rely on the default model and switch later if needed. Beyond text, the app supports voice input, converting speech to text in the prompt box so users can think out loud and edit before sending. The tool also offers session-level customization: users can select different agents configured for different types of work, and can enable remote control to access the same session from the web, allowing work to be paused on one device and resumed on another. The article stresses that none of these options require pre-configuration; they are available when needed. The core recommendation is to start with a small change in a project you already know, describe it in plain language, and iterate from there.