For AstraZeneca, migrating to GitHub wasn’t just about moving where their code lives—it was about transforming how teams use software to accelerate the discovery and delivery of life-saving medicine. When they migrated to GitHub Enterprise, they standardized their developer environments, automated manual workflows with GitHub Actions, and, with the peace of mind of GitHub’s industry leading duplication detection filter, began leveraging agentic AI with GitHub Copilot. These changes reduced friction and helped them turn complex challenges into tasks that they could accomplish in hours instead of weeks, giving engineers valuable time to refocus on their mission.

In the early results from AstraZeneca’s GitHub Copilot pilot program, developer velocity increased by 40%—a direct acceleration for the engineering teams who support the organization’s most critical work, like AI-powered drug discovery and global clinical trials. With a drug development lifecycle that can take over a decade, shaving even one year off that timeline can have a massive impact on the business and patient outcomes.

Building a strong foundation

Before adopting GitHub, AstraZeneca’s software teams were using a mix of tools and processes that complicated collaboration and made it difficult to scale. With a diverse user base of around 3,000 people accessing source code, from traditional software engineers to data scientists and researchers, the company saw an opportunity for a more unified, modern platform to streamline workflows, enhance collaboration, and reduce risk. Their ideal solution would need to be both intuitive for developers and technically robust for automated deployments.

The adoption of industry-standard platforms like GitHub has significantly streamlined our onboarding process. New engineers immediately recognize the value of working with familiar, cutting-edge development tools.

GitHub’s widespread adoption meant most engineers would already be familiar with the platform, so the onboarding experience would be intuitive, and the ramp-up for new hires would be faster, too. 

According to Paul Ordman, Director of Software Engineering at AstraZeneca, "The adoption of industry-standard platforms like GitHub has significantly streamlined our onboarding process. New engineers immediately recognize the value of working with familiar, cutting-edge development tools."

More than just a familiar interface, GitHub’s built-in CI/CD capabilities were also a major draw. AstraZeneca wanted to shift left, bringing automated testing, builds, and deployments into a single, reliable system to directly address their inconsistent and error-prone deployments. “We wanted easier CI/CD tools to automate the process of building and releasing software,” Ordman said. “To make it reliable and consistent.”

Unifying 3,000 developers on a single platform

The first step in AstraZeneca’s transformation was a monumental one: consolidating thousands of developers from a fragmented landscape of tools, including Bitbucket and other satellite systems, onto a single, unified platform. The stakes of such a large-scale project—moving approximately 3,000 users and 20,000 repositories—were high, with no room for error or interruption to business-critical work. 

The process required a hybrid approach of GitHub's native capabilities and AstraZeneca’s ingenuity. For the technical heavy lifting of moving repositories and recreating permissions, they used GitHub Issues and Actions; and to manage the human side of migration, they built custom software, which automated the complex outreach, data collection, and scheduling for every team, ensuring a predictable and well-managed process for all 3,000 engineers, data scientists, and researchers.

The migration was a carefully planned strategic initiative, involving a full year of preparation followed by a meticulous four-month execution phase. By the end of the process, AstraZeneca had successfully achieved its goal. Nearly 99% of the company's developers and source code were consolidated on GitHub, setting a stable, unified foundation for the next phase of their transformation.

Success breeds success: the initial 3,000 migrated users inspired 2,000 more developers and data scientists to adopt the platform by the end of the year. 

Unlocking value with automation

The returns on this migration investment began with having their code all in one place, and continued with AstraZeneca’s new automation capabilities unlocked by the platform. In the months following the migration, the number of automated CI/CD events running on the platform grew by 100% and are still growing rapidly. 

The ease of use and robust functionality of GitHub Actions has been a true game changer for scaling automation across our entire organization.

"Every automated CI/CD pipeline that executes represents value and efficiency," Ordman explained. "The ease of use and robust functionality of GitHub Actions has been a true game changer for scaling automation across our entire organization."

This widespread automation enabled AstraZeneca's teams to successfully shift left. Instead of finding bugs weeks after code was written, teams can now find and fix defects in hours. The impact on the business has been profound. According to Ordman, identifying issues at this early stage results in "exponentially reduced remediation costs—often tens or hundreds of times more cost-effective than addressing problems in later development phases."

AstraZeneca's approach to scaling this success was guided by a simple philosophy from Ordman: “How do we make it easy for teams to do the right thing?” 

To put that philosophy into practice, the central DevOps team at AstraZeneca created a library of reusable GitHub Actions for common, critical tasks like security scanning and automated testing. These actions were designed to function like a “black box,” Ordman noted. Teams don't need to understand the complex code inside; they simply provide a few parameters to get a reliable and consistent outcome. This approach scaled best practices across the organization and empower every team to build automation into their workflow with minimal effort.

Supercharging developers with AI

AstraZeneca next turned to AI to increase developer productivity. For a company in a highly regulated industry, adopting GitHub Copilot required a thoughtful approach that balanced innovation with robust governance. The journey began only after GitHub's crucial legal indemnification for Copilot became available for enterprise customers, a prerequisite for protecting AstraZeneca's valuable intellectual property. Before any developer had access, the tool also underwent a full AI, cybersecurity, and ethics review.

Once these safeguards were in place, the rollout began, and the demand from developers was immediate and overwhelming. A goal to onboard 1,000 users within a year was shattered in just six months, with over 1,600 developers and scientists onboarded, and active usage growing by 20% per month. 

To unlock Copilot's potential and address developer concerns about the impact of AI on their roles, AstraZeneca found success with a hands-on, team-based training approach. The impact was big: by learning to use Copilot’s agent mode in their daily work, pilot squads saw an immediate 40% increase in development velocity. This velocity boost is reflected in developer-reported gains as well: an internal survey found that engineers are getting 9-10 hours of extra output per week—a significant jump from four hours reported in the previous quarter, which Ordman’s teams credit to the adoption of Copilot's agent mode.

GitHub Copilot delivered immediate impact. Achieving return on investment within a single sprint cycle rarely occurs with technology implementations.

"GitHub Copilot delivered immediate impact," Ordman stated. "Achieving return on investment within a single sprint cycle rarely occurs with technology implementations."

While this impressive productivity boost comes from the tool's overall daily usage, Ordman points to Copilot Agent Mode and early usage of Copilot coding agent, which is designed for more complex, end-to-end tasks, as the source of the most dramatic productivity lifts. For instance, a developer used the agent to complete a project in just two hours that would have previously taken two developers a full week. 

Yet, for all the speed AI unlocks, AstraZeneca’s culture of rigorous quality assurance remains unchanged. The company is clear that AI is a powerful accelerator, not a replacement for expertise. “You still need skilled engineers to understand every line and every nuance of why the system behaves the way it does,” Ordman said. This philosophy extends to their processes: In an industry where precision is mission-critical, their approach to validation is unwavering. Ordman put it simply, “The methods change—the mission of testing does not.”

The result: a world-class engineering platform

From a fragmented collection of tools to a unified platform powered by automation and AI, AstraZeneca has fundamentally reshaped its software development lifecycle. The impact goes beyond developer efficiency, directly accelerating the speed at which the company can deliver on its mission of providing life-saving medicine. 

For Ordman, the results of this approach—leveraging the platform for powerful automation and AI—are unmistakable. “The combination of GitHub Actions and Copilot has been transformative.” he said. “It's a night and day difference for our developer and team productivity, and GitHub is only accelerating the release of new features, so I’m expecting to see continuous productivity improvements in the years ahead.” With a world-class engineering platform now in place, AstraZeneca is positioned to innovate faster than ever.