GitHub Copilot can now turn Microsoft Teams notifications into pull requests

Bridging communication and code

GitHub has introduced a new update to its Microsoft Teams app that integrates AI-driven coding capabilities directly into team communication. This update allows developers to interact with GitHub Copilot to investigate repository updates and prepare code changes without leaving their chat environment. By connecting notifications with Copilot, the platform aims to streamline how teams move from identifying a problem to implementing a solution.

The primary goal of this feature is to reduce the friction often found in software development workflows. Instead of switching between a chat app and a code editor, developers can now handle tasks directly from notification threads. This integration ensures that the conversation surrounding a bug or a new feature stays closely tied to the actual work being performed on the repository.

As Analytics Insight reported, the update allows users to reply to a repository update and mention @GitHub to initiate a task. By describing the work they want completed, users can prompt the AI to take action based on the specific context of the discussion. This creates a more fluid transition from high-level planning to technical execution within a single interface.

How the AI handles the workflow

When a user triggers Copilot through a Teams thread, the AI uses two primary sources of information to understand the request. It looks at the linked issue or pull request, as well as the ongoing conversation in the Teams thread. This dual context allows Copilot to understand the nuances of the task as discussed by the human team members.

The AI is capable of performing several complex technical actions autonomously. It can reproduce bugs to verify their existence, run tests to ensure stability, and make the necessary code fixes. Once the work is complete, Copilot can submit these changes as a proposed pull request for the team to evaluate.

Beyond just fixing bugs, teams can use this same workflow for broader development goals. This includes adding new tests to an existing codebase or updating technical documentation. Because Copilot uses the existing discussion as context, teams can avoid the tedious process of rewriting detailed task descriptions when moving from a conversation to a formal implementation.

Access and administrative requirements

While the ability to receive GitHub notifications does not require a specific license, using the Copilot features does. Users must have a paid GitHub Copilot plan that includes enabled cloud sandboxes to use these capabilities. Cloud sandboxes are isolated environments where the AI can safely run code and tests without affecting the main production environment.

For users within larger companies, there may be additional hurdles to getting started. Organization and enterprise users might need an administrator to specifically enable both the Copilot cloud agent and the cloud sandboxes. These permissions are necessary to allow the AI to delegate work and perform actions through the Teams interface on behalf of the user.

Existing users can access these new capabilities by simply updating their current app. However, those using the legacy GitHub app will need to take an extra step. They must install the consolidated GitHub app to access Copilot through Teams, while the older version will be renamed to 'GitHub (deprecated)' until its retirement date.

The path toward agentic fixes

This update is part of a broader trend toward 'agentic' workflows in software development. GitHub has already made Agentic autofix generally available for code scanning alerts. This system allows the AI to automatically suggest fixes for security or quality issues, which are then checked through CodeQL before reaching a human reviewer.

The integration with Microsoft Teams follows a similar logic of moving toward automated, context-aware assistance. By allowing the AI to act on notifications, GitHub is positioning Copilot as a proactive agent rather than just a passive autocomplete tool. This shift is designed to help developers spend less time on repetitive manual tasks and more time on high-level problem solving.

Microsoft product and engineering teams are already utilizing the Teams app to turn GitHub issues into code. This internal adoption serves as a blueprint for how broader teams can use the tool to manage their own repositories. As the technology matures, the goal is to make the transition from a chat message to a merged pull request as seamless as possible.

Timeline and app migration

Users who want to stay current with these features should look for the consolidated GitHub app in the Microsoft Teams marketplace. This app serves as the primary hub for both standard notifications and the new Copilot-powered features. It is the only version that will support the advanced AI capabilities described in the update.

The transition period for the old app is relatively long, providing users plenty of time to migrate their workflows. The legacy app is scheduled to retire on November 16, 2026. Until that date, it will remain available but will be clearly marked as deprecated to encourage users to move to the newer, more capable version.

Once migrated, users can follow prompts to connect their GitHub accounts and select specific repositories. This setup ensures that the AI has the correct permissions to interact with the intended code. By centralizing these tools, GitHub aims to provide a unified experience for developers who rely on Teams as their primary communication hub.

Next steps for teams

Teams looking to adopt this workflow should first verify their licensing and administrative settings. Ensuring that cloud sandboxes are active is a critical first step for any organization wanting to let Copilot perform actions like running tests or submitting pull requests. Without these permissions, the AI will be limited in what it can do autonomously.

Developers can start by experimenting with the @GitHub mention in their existing notification threads. By describing simple tasks, such as updating a piece of documentation or creating a basic test, teams can get a feel for how well the AI interprets the conversation context. This helps establish a baseline for how the team wants to interact with the AI agent.

As the feature rolls out, it will be important to monitor the quality of the AI-generated pull requests. While Copilot can handle many routine tasks, human review remains a necessary step in the process. Teams should decide on a standard review cadence to ensure that the AI-generated fixes meet their specific coding standards and security requirements.

Try it

Check your Microsoft Teams app list to see if you have the consolidated GitHub app installed, or look for the 'GitHub (deprecated)' version to plan your migration.

Watch

How many organizations will choose to enable cloud sandboxes for Copilot.

Whether GitHub will introduce more 'agentic' features for non-coding tasks like project management.

The accuracy of Copilot when interpreting complex, multi-turn conversations in Teams threads.