GitHub Copilot’s Slack and Microsoft Teams integrations now accept richer conversation artifacts, let users lock in model choices and default repositories, and include reliability fixes for longer‑running tasks. This matters because engineers can generate code, issues, and implementation plans directly from the context already present in their chat channels, cutting down on context‑switching and preserving traceability.
Richer Context from Chat Artifacts
In Slack, Copilot can now ingest attached files, uploaded documents, and direct links to messages as part of the prompt. In Teams, the agent can process inline images, forwarded messages, and the full history of a channel or thread. Before opening a new issue, Copilot checks for existing similar issues, attaches a direct link to the created work item, and maintains a backlink to the originating conversation, making the decision trail visible to the whole team.
More Control Over Work Generation
Users may select a specific AI model for the next message and have that choice persist throughout the conversation. Slack adds the ability to set default owners and target repositories, which streamlines work that spans multiple repos and shared discussions. Repository switching has been hardened so that a superseded session cannot continue acting on an old repository, reducing accidental cross‑repo actions.
Reliability and Operational Improvements
The update improves handling of longer‑running tasks by providing clearer implementation‑plan status and more predictable reconnection when a conversation goes idle. In Teams, the agent now retains channel‑thread history more reliably, avoids duplicate answers, correctly processes Teams‑converted images, and works consistently with user‑owned repositories and large channels. Slack sees fixes to the implementation‑plan recovery flow, repository picker, and code‑channel interactions, along with safer repository switching.
Availability and Prerequisites
The enhancements are in public preview for organizations on GitHub Copilot Business and Enterprise plans, counting against existing Copilot entitlements and managed through existing cloud‑agent budgets. Administrators must enable the Copilot cloud agent policy (Slack) or the cloud agent and cloud sandboxes (Teams) before users can install or upgrade the respective GitHub apps. After installation, users mention @GitHub in a conversation to trigger the agent.
Related CloudNinjas coverage: AI engineering.
What This Means For Practitioners
- Evaluate whether the richer context handling aligns with your team’s workflow for sharing artifacts in chat; adjust channel policies if needed.
- Review default owner and repository settings to ensure they match your governance model and avoid unintended cross‑repo commits.
- Monitor the implementation‑plan status messages for longer tasks; integrate them into existing observability pipelines if you rely on automated alerts.
- Confirm that the required admin policies (cloud agent, sandboxes) are in place and that usage is tracked against your Copilot entitlement budget.
- Test the new model‑selection persistence in a staging channel to verify that the chosen model behaves as expected across conversation turns.