Module (08): (GHDR) Developer Use Cases for AI with GitHub Copilot

[MohamedRadwan-DevOps]

Module (08): (GHDR) Developer Use Cases for AI with GitHub Copilot

Document Type: GHDR (GitHub Documentation Reference)
Scope: This document provides the official extracted content from GitHub Docs corresponding to Module (08): Developer Use Cases for AI with GitHub Copilot. It preserves the original wording exactly as published, extending and complementing the Microsoft Learn module with additional verified information.

Introduction

Where you can use Copilot

Copilot is available:

• In your IDE
• In GitHub Mobile, as a chat interface
• In Windows Terminal Canary, through the Terminal Chat interface
• On the command line, through the GitHub CLI
• On the GitHub website

Source: What is GitHub Copilot?

Getting access to Copilot

For individuals

• Try Copilot for free: Use Copilot Free to explore core Copilot features with no paid plan required.
• Subscribe to a paid plan: Upgrade to Copilot Pro or Copilot Pro+ for full access to premium features and more generous usage limits. You can try Copilot Pro for free with a one-time 30-day trial.
• Eligible for free Copilot Pro access? Students, teachers, and open-source maintainers may qualify for Copilot Pro at no cost.
• Organization members: If your organization or enterprise has a GitHub Copilot plan, you can request access by going to https://github.com/settings/copilot and requesting access under “Get Copilot from an organization.”

For organizations and enterprises

• Organization owners: Purchase Copilot Business for your team.
• Enterprise owners: Purchase Copilot Business or Copilot Enterprise for your enterprise.
• If you don’t need other GitHub features, you can create an enterprise account specifically for managing Copilot Business licenses. This gives you enterprise-grade authentication options without charges for GitHub Enterprise licenses.

Source: What is GitHub Copilot?

Choosing the right Copilot interface

While Copilot inline suggestions and Copilot Chat share some functionality, the two tools are best used in different circumstances.

Inline suggestions work best for:

• Completing code snippets, variable names, and functions as you write them
• Generating repetitive code
• Generating code from inline comments in natural language
• Generating tests for test-driven development

Copilot Chat is best suited for:

• Answering questions about code in natural language
• Generating large sections of code, then iterating on that code to meet your needs
• Accomplishing specific tasks with keywords and skills
• Completing a task as a specific persona

Source: Best Practices for Using GitHub Copilot

Validate and improve Copilot outputs

While Copilot is very powerful, it is still a tool capable of making mistakes, and you should always validate the code it suggests. Recommended actions include:

• Understand suggested code before you implement it
• Review Copilot’s suggestions carefully
• Use automated tests and tooling to check Copilot’s work
• (Optional) You may want to check Copilot’s work for similarities to existing public code and turn off suggestions matching public code if desired

Source: Best Practices for Using GitHub Copilot

Stay current

New features are regularly added to Copilot to create new abilities, build on existing features, and improve the user experience. To stay up-to-date with Copilot’s features, see the changelog.

Source: Best Practices for Using GitHub Copilot

Boost developer productivity with AI

Choosing your enterprise’s plan for GitHub Copilot

• “GitHub offers two Copilot plans for customers on GitHub Enterprise Cloud:
  Copilot Business ($19 USD per user per month) … Copilot Enterprise ($39 USD per user per month) …”
• “Premium requests are used by more advanced Copilot features and models, including AI agents.”
• “By default, users can continue to use premium requests after exhausting their plan’s monthly allowance. However, each extra request is charged to your enterprise…”
• “Copilot Business users who use more than around 800 premium requests a month would save money on a Copilot Enterprise plan.”

Source: Choosing your enterprise’s plan for GitHub Copilot

Requests in GitHub Copilot (premium requests)

• “Billing for premium requests began on June 18, 2025 … and on August 1, 2025 on GHE.com.”
• “Premium request counters reset on the 1st of each month at 00:00:00 UTC.”
• “If you use Copilot Free, your plan comes with up to 2,000 inline suggestion requests and up to 50 premium requests per month.”
• “If you’re on a paid plan, you get unlimited inline suggestions and unlimited chat interactions using the included models … Paid plans also receive a monthly allowance of premium requests.”
• “GPT-5 mini, GPT-4.1 and GPT-4o are the included models, and do not consume any premium requests if you are on a paid plan.”
• “Each model has a premium request multiplier … e.g., Claude Opus 4.1 = 10×; GPT-4.1/GPT-4o (paid) = 0×.”

Source: Requests in GitHub Copilot

Getting code suggestions in your IDE

• “GitHub Copilot offers coding suggestions as you type.”
• Accept suggestions with Tab; show alternatives with Alt+] / Alt+[ (Windows/Linux) or ⌥] / ⌥[ (macOS).
• “Press Ctrl+Enter to open a new tab with multiple additional options.”
• “You can accept the next word of a suggestion (macOS ⌘→ / Windows/Linux Ctrl→).”
• “Next edit suggestions predict where and what edits may be needed … navigate using Tab; a gutter arrow indicates an available edit.”

Source: Getting code suggestions in your IDE with GitHub Copilot

Syncing documentation with code changes

• “Copilot Chat can assist in updating existing code documentation.”
• Example shows updating function docs to include new parameters and error cases so comments match the current implementation.

Source: Syncing documentation with code changes

Using GitHub Copilot to explore projects

• In GitHub, you can “Chat about this repository,” ask “Summarize the purpose of this repository based on the README,” or request explanations for selected lines.
• File view offers a Copilot icon to ask about the entire file; for ranges, select lines and use the inline Copilot button or predefined questions.

Source: Using GitHub Copilot to explore projects

Copilot text completion (PR descriptions) — public preview

• “Use AI-generated autocompletions to help you write pull request descriptions quickly and accurately.”
• “To participate in the public preview, an administrator must opt in to the use of previews of Copilot features.”

Source: GitHub Copilot text completion

Creating a pull request summary with GitHub Copilot

• “You can generate a summary in the description of a pull request, or as a comment.”
• “Copilot does not take into account any existing content in the PR description, so it is best to start with a blank description.”
• Use the Copilot icon → Summary, review, and optionally rate with thumbs up/down.

Source: Creating a pull request summary with GitHub Copilot

Using GitHub Copilot code review

• “Open the Reviewers menu, then select Copilot.”
• “Copilot always leaves a ‘Comment’ review, not ‘Approve’ or ‘Request changes’ … does not count toward required approvals.”
• “Where possible, Copilot’s feedback includes suggested changes … you can accept individually or as a group.”
• You can enable automatic reviews and add repository custom instructions (e.g., .github/copilot-instructions.md).

Source: Using GitHub Copilot code review

About GitHub Copilot coding agent

• “Copilot coding agent can: Fix bugs; Implement incremental new features; Improve test coverage; Update documentation; Address technical debt.”
• Works “independently in the background … makes the required changes and opens a pull request … then requests a review from you.”
• Runs in an ephemeral GitHub Actions environment with tests/linters; supports custom agents for specialized tasks.
• Security protections include CodeQL scanning, advisory/vulnerability checks, secret scanning, restricted branches (copilot/*), sandboxed environment, and compliance/audit controls.
• Limitations: single-repo scope, one PR per task, respects governance, no model picker (currently Claude Sonnet 4.5), content exclusions not applied, GitHub-hosted only.

Source: About GitHub Copilot coding agent

Prompt engineering for GitHub Copilot Chat

• “Start general, then get specific,” “Give examples,” “Break complex tasks into simpler tasks,” and “Avoid ambiguity.”
• “Indicate relevant code” by opening files / selecting code; use participants like @workspace (VS Code) or @project (JetBrains).
• “Experiment and iterate,” keep history relevant (threads, delete irrelevant turns), and follow good coding practices.

Source: Prompt engineering for GitHub Copilot Chat

Align with developer preferences

Personal custom instructions

• “Personal custom instructions are currently only supported for GitHub Copilot Chat in GitHub.”
• “You can add custom instructions … to receive chat responses that are customized to your personal preferences.” Examples: “Always respond in Spanish.” / “Always provide examples in TypeScript.”
• Priority: “Personal instructions take the highest priority, followed by repository instructions, with organization instructions prioritized last.”
• How to add: “Open Copilot Chat… click your profile picture → Personal instructions… add natural language instructions… click Save.”

Source: Adding personal custom instructions for GitHub Copilot

Repository custom instructions

• Types: “Repository-wide custom instructions” in .github/copilot-instructions.md; “Path-specific custom instructions” in .github/instructions/NAME.instructions.md using applyTo globs; “Agent instructions” via AGENTS.md (or CLAUDE.md / GEMINI.md).
• Example frontmatter:

  ---
  applyTo: "**/*.ts,**/*.tsx"
  ---
  

• Optional: excludeAgent: "code-review" (or "coding-agent").
• Usage: “The complete set of instructions will be automatically added to requests… In Copilot Chat’s immersive view… the instructions file is added as a reference for the response.”
• Enable/disable: In a repo with instructions, Chat panel → … → Disable/Enable custom instructions; for code review: Settings → Copilot → Code review → Use custom instructions…

Source: Adding repository custom instructions for GitHub Copilot

Organization custom instructions

• Availability: “Organization owners… (Copilot Business or Copilot Enterprise).” “This feature is currently in public preview.”
• Surfaces: “Supported for Copilot Chat on GitHub.com, Copilot code review on GitHub.com and Copilot coding agent on GitHub.com.”
• How to add: Org Settings → Copilot → Custom instructions → Preferences and instructions → Save changes.

Source: Adding organization custom instructions for GitHub Copilot

AI model comparison (choose the best model for your task)

• “GitHub Copilot supports multiple AI models with different capabilities.” “Different models have different premium request multipliers.”
• Auto-select: “When you use Copilot Chat in supported IDEs, Auto will automatically select the best model… You can manually choose a different model.”
• Promotion: “Complimentary access to Grok Code Fast 1 is continuing past the previously announced end time.”
• Task guidance (examples):
  • General-purpose coding and writing: “GPT-5 mini… reliable default”; “Grok Code Fast 1… specialized for coding tasks”; “Raptor mini… specialized for fast, accurate inline suggestions.”
  • Fast help with simple or repetitive tasks: “Claude Haiku 4.5… Ideal for small tasks.”
  • Deep reasoning and debugging: “GPT-5 mini… deep reasoning and debugging… faster responses”; “GPT-5… complex reasoning”; “Claude Sonnet 4… smarter reasoning”; “Gemini 2.5 Pro… advanced reasoning across long contexts.”
  • Working with visuals: “GPT-5 mini… supports multimodal input”; “Gemini 2.5 Pro… visual reasoning.”

Source: AI model comparison

Changing the AI model for Copilot Chat

• “Different models have different premium request multipliers.”
• “Copilot allows you to change the model during a chat…” (Business orgs must allow switching.)
• Limitation: “You can only use an alternative AI model in the immersive view… The Copilot Chat panel always uses the default model.”
• Steps (GitHub website): Copilot button → immersive view → CURRENT-MODEL dropdown → choose model; optional retry to regenerate with the new model.

Source: Changing the AI model for GitHub Copilot Chat

Changing the AI model for inline suggestions

• Prereqs: “An alternative model is currently available” and latest VS Code + Copilot extension.
• Steps (VS Code): Command Palette → “GitHub Copilot: Change Completions Model” → select model.
• Check current model: Settings → “GitHub > Copilot: Selected Completion Model.”

Source: Changing the AI model for GitHub Copilot inline suggestions

Examples of task-aligned assistance

Debugging invalid JSON

• “Copilot Chat can identify and resolve syntax errors or structural issues in JSON data.” (Example: missing closing quote causing parse error; shows corrected JSON.)

Source: Debugging invalid JSON

Handling API rate limits

• “Copilot Chat can help handle API rate limits by suggesting code that detects and implements retry logic.” (Example: exponential backoff with requests + HTTPAdapter/Retry.)

Source: Handling API rate limits

AI in the Software Development Lifecycle (SDLC)

Plan & Track (Issues)

• Create or update issues from natural language (or screenshots) in the Copilot Chat immersive view. Copilot drafts the title, body, labels, assignees, and respects your repo’s issue forms/templates. You can edit, switch templates without losing content, or ask Copilot to revise. It also supports multiple issues in one prompt and sub‑issues (issue trees); you can review an issue tree, expand/collapse items, and publish each draft. Permissions are unchanged—you must already be allowed to create issues in the target repo.

• Assign issues to Copilot (requires Copilot coding agent enabled). Ask naturally (“Assign this issue to Copilot”) or choose Copilot in assignees; Copilot starts a session and opens a PR with its changes.

Source: Using GitHub Copilot to create or update issues

Develop & Review (Pull Requests)

• Explore pull requests with Copilot Chat: from a PR you can ask for a summary, current status, or the main changes. On Files changed, you can ask about a specific file or selected lines via the Copilot button. Copilot can also explain why a workflow failed from the PR’s failing check details page and suggest fixes.

Source: Using GitHub Copilot to explore pull requests

CI/CD (GitHub Actions)

• Open Copilot from failed runs: in a PR’s merge box, use Explain error to open a Copilot chat about the failed workflow; Copilot provides steps to resolve. Use workflow logs, debug logging, and review billing/budgets when failures relate to usage. Includes guidance on triggers, filters & diff limits (first 300 files for path filters), scheduled workflow delays, and force‑cancelling runs via REST.

Source: Troubleshooting workflows

Secure (Code Scanning & Autofix)

• Copilot Autofix can generate fixes for supported CodeQL alerts: open an alert → Generate fix → Create PR with fix (draft PR on a new branch). Also available via Autofix historical alerts API.

• Assign alerts to Copilot coding agent (public preview): from an alert, choose Assignees → Copilot; Copilot opens a PR within ~30 s with a summary and changes.

• Dismiss, bulk‑dismiss, or re‑open alerts (records reason; removal applies across branches). Manage stale configurations per branch from the alert’s Affected branches panel.

Source: Resolving code scanning alerts

Understand limitations and measure impact

Rate limits for GitHub Copilot

• GitHub enforces rate limits for capacity, fairness, and abuse mitigation.
• When you exceed limits, you may temporarily lose access to certain Copilot features or models.
• If you receive a rate-limit error:
  • Wait and try again — limits are temporary.
  • Check and reduce rapid or automated requests.
  • Switch to another model (preview models often have stricter limits).
  • Contact GitHub Support if legitimate use is affected.

Source: Rate limits for GitHub Copilot

Data available in Copilot usage metrics

• The Copilot usage metrics dashboard and APIs display a consistent set of data fields, including:
  • Agent adoption, daily/weekly active users, acceptance rate, model usage, and language usage.
  • code_generation_activity_count, code_acceptance_activity_count, and loc_suggested_to_add_sum for granular reporting.
• Data may appear up to three UTC days delayed, based on IDE telemetry.
• Exports are available in NDJSON format and can be analyzed in Copilot Chat with natural-language queries.

Source: Data available in Copilot usage metrics

REST API endpoints for Copilot metrics

• Use Copilot Metrics REST API to retrieve enterprise, organization, or team metrics.
• Endpoints include:
  • /orgs/{org}/copilot/metrics – Organization metrics.
  • /enterprises/{enterprise}/copilot/metrics – Enterprise metrics.
  • /enterprises/{enterprise}/team/{team_slug}/copilot/metrics – Team metrics.
• Supports parameters like since, until, page, and per_page (up to 100 days of data).
• Tokens require manage_billing:copilot, read:org, or read:enterprise scopes.
• API returns metrics for active and engaged users, languages, and IDEs, with model-specific breakdowns.

Source: REST API endpoints for Copilot metrics

Reconciling Copilot usage metrics across dashboards, APIs, and reports

• All data sources share telemetry but differ by aggregation window and processing timing:
  • Dashboard: 28-day rolling average.
  • APIs and exports: Daily records for trend analysis.
• Discrepancies may occur due to delayed telemetry (≈ 3 days), export timing, or Unknown values from older IDE clients.
• Aligning report windows with the dashboard’s 28-day period yields the most consistent results.

Source: Reconciling Copilot usage metrics across dashboards, APIs, and reports

Monitoring your GitHub Copilot usage and entitlements

• You can track monthly premium request usage in multiple ways:
  • Directly within your IDE (status-bar Copilot icon).
  • In Billing and Licensing at https://github.com/settings/billing.
  • Under Premium request analytics to view or download charts.
• Counters reset monthly (UTC 00:00 on the 1st).
• Admins can view aggregated usage for organizations and enterprises.
• Optimization tips:
  • Choose the right model for each task.
  • Set budgets and receive usage alerts (75 % / 90 % / 100 %).
  • Avoid retrying large prompts unnecessarily.
  • Manage whether users can exceed allowances through policy.

Source: Monitoring your GitHub Copilot usage and entitlements