AI is changing how software gets planned, built, and reviewed. As teams adopt agentic development, the platform underneath those workflows matters more. They need tools that bring planning, coding, security, and collaboration together—and can keep pace with how development is evolving.
That’s why we’re delivering the newest agentic capabilities on GitHub across planning, coding, code review, and security. For teams driving active development, that often means moving repositories to GitHub to unlock the latest AI-powered workflows, while continuing to use Azure Boards and Pipelines. For teams that need more time migrating repos to GitHub, we’re continuing to invest in Azure DevOps with improvements focused on security and code quality in the workflows they rely on today.
Unlocking agentic development on GitHub
Organizations that move repositories to GitHub unlock the full agentic development experience, including a broad choice of models from Anthropic, OpenAI, Google, and others in Copilot, along with agentic workflows across GitHub, VS Code, mobile, CLI, and integrations like Teams, Slack, and Azure Boards. Enterprise governance is built in through an agent control plane that provides visibility, audit, and policy management.
Teams can continue using Azure DevOps capabilities such as Azure Boards, Pipelines, and Test Plans alongside their GitHub repos. Azure DevOps basic usage rights are included with GitHub Enterprise, making it easier to adopt a hybrid model without additional overhead. Enterprise Live Migrations and deeper GitHub–Azure DevOps integration make it practical to transition at enterprise scale, as reflected in Microsoft’s own migration journey.
Enterprise Live Migrations — Preview
Large repos and complex organizations need migration options that respect how their teams operate. Enterprise Live Migrations (ELM), now available in preview, enables organizations to migrate multiple repositories together with minimal cutover downtime. ELM supports starting migration without locking the Azure DevOps repository, allowing developers to continue working while teams transition at a pace that fits their portfolio and operational needs. Repository history, branches, and metadata are preserved throughout the migration process. For organizations choosing GitHub for source control while continuing to use Azure DevOps for planning, CI/CD and testing, ELM is designed to make that transition significantly easier and less disruptive. A script-based migration experience is available today in preview, with a more streamlined end-to-end experience coming soon. Join the waitlist for early access to Enterprise Live Migration.
Customer zero: How Microsoft is moving to GitHub internally
Microsoft is using this model internally too. In the Copilot, Agents, and Platform team, a small migration team has moved 1,575 repositories from Azure Repos to GitHub Enterprise, helping drive a company-wide shift at scale. Today, more than 3,000 developers are on GitHub, and 45% of pull requests now happen there.
AI is the forcing function. GitHub is where Copilot, agents, and new AI workflows ship first and run at scale, so moving source control there gives teams faster access to the latest capabilities. At the same time, teams can continue to use Azure Pipelines and Azure Boards in Azure DevOps while they move their code to GitHub, without waiting for a full platform transformation.
For a deeper look at how the team approached the migration, including practical best practices and lessons learned, see the related blog post.
Connecting agents to your DevOps context
As teams adopt hybrid workflows across GitHub and Azure DevOps, agents need context from the systems where work gets planned, built, reviewed, and shipped. GitHub already provides a remote MCP server for repository and workflow context. The new Azure DevOps remote MCP server in preview extends that model to Azure DevOps, bringing work items, builds, pull requests, test plans, and other project context directly into agentic workflows.
The service is fully hosted by Microsoft, geo-routed, and stateless, with no Azure DevOps customer data persisted within the MCP service itself. It’s also available in Azure AI Foundry for building custom agents, and we’re working to make it available in Copilot Studio and Microsoft 365 Copilot so teams can connect Azure DevOps to their agents without managing additional infrastructure.
Security and quality improvements for Azure DevOps repositories
For teams continuing to run repos on Azure DevOps, we’re bringing AI-powered capabilities directly into the workflows they already use. The updates below focus on two areas that matter most: improving code quality in pull requests and helping developers remediate security issues faster.
Copilot Code Review for Azure DevOps — Preview
With Copilot Code Review, now in preview, developers can request an automated review on any pull request and receive inline comments and suggestions directly in the Azure DevOps pull request experience. Organization administrators maintain full control over adoption and enablement through settings at the org, project, or repository level. Usage is billed using GitHub AI Credits through the same Azure subscription your organization already uses for Azure DevOps, with no separate license management required. Copilot Code Review is designed to complement, not replace, human reviewers. It doesn’t block merges or count toward required reviewers. Instead, it acts as an additional layer of review that helps surface issues earlier, allowing engineers to focus their attention on the changes that matter most. Join the waitlist for early access to Copilot Code Review for Azure DevOps.
Autofix for CodeQL in Azure DevOps — Preview
Security issues are only valuable to surface if teams can fix them efficiently. Autofix for CodeQL alerts, now in preview for Azure Repos as part of GitHub Advanced Security for Azure DevOps, brings AI-powered remediation directly into the code security workflow. With Autofix, developers can request a suggested fix for an open CodeQL alert, and Copilot automatically creates a pull request with the proposed remediation ready for review. It’s the same Autofix experience GitHub customers have used to significantly reduce time-to-remediation for security issues. Preview today covers a subset of CodeQL alerts and will expand over time toward broader coverage across code scanning alerts as we move toward general availability. Join the waitlist for early access to autofix for CodeQL in Azure DevOps.
Apple Silicon Macs for Azure Pipelines, with pay-per-minute billing
Apple Silicon Macs — a capability many customers have been waiting for — are now available as Microsoft-hosted agents in Azure Pipelines with pay-per-minute billing. Teams can now run faster Apple-native builds for iOS and macOS without managing their own Mac infrastructure or planning capacity for parallel jobs, while paying through a flexible usage-based model that scales with pipeline demand.
Moving forward
GitHub is where the newest AI-powered development workflows are taking shape, and for many teams that will increasingly mean moving source control there over time. We’re investing to make that transition practical, whether you’re migrating repositories now, adopting a hybrid Azure DevOps + GitHub model, or moving repositories in phases based on complexity and business priorities.
Across each of those paths, the goal is the same: help teams move forward with stronger security, better code quality, and access to the latest agentic capabilities.
The post Azure DevOps and GitHub: Journeying into the AI Era appeared first on Azure DevOps Blog.