How Thomson Reuters successfully adopted AI (and how your organization can, too)

Has your organization implemented AI into your software engineering process yet? 

Across industries, organizations are turning to AI to automate repetitive tasks, streamline complex processes, and enable workers to focus on more strategic, creative initiatives. Tools like GitHub Copilot—shown to speed up task completion by 46% and improve code quality by 39%—are already transforming the way software gets built.

But successful AI adoption is not so simple. It demands strategic planning, structured implementation, and aligning the technology with organizational goals. Businesses large and small will have to implement an individualized and mindful approach to get the most out of an AI solution. 

One company that has had success with GitHub Copilot adoption is Thomson Reuters, a legal, tax, accounting, and media leader. Let’s take a look at the steps they used to implement and scale GitHub Copilot across their engineering organization, and how you can do the same.

Setting the stage

Thomson Reuters began their AI journey in 2022. Engineering leaders realized that what started as industry buzz was an opportunity to foster engineering excellence within the company–and GitHub Copilot could be just the thing to accelerate their operations. 

But first, they needed to answer several key questions:

• How can GitHub Copilot improve developer efficiency, code quality, and satisfaction?
• Will it align with Thomson Reuters’ broader engineering culture and business goals?
• How can they onboard developers effectively with minimal resistance?

To find these answers, they initiated two small, focused GitHub Copilot pilot programs in  2023. The programs included:

• 100+ engineers from various teams working across a range of tech stacks.
• A commitment from participants to engage in four hours of coding per week with GitHub Copilot and provide weekly feedback.
• A seven-week timeline designed to thoroughly test how GitHub Copilot impacted coding efficiency, quality, and the overall developer experience.

Pilot program results

How did it go? 

• 46% faster task completion rate for developers using GitHub Copilot.
• 39% improvement in code quality (measured by the number of changes per PR, indicating more impactful contributions).
• 68% of developers reported a positive user experience, describing the tool as intuitive and easy to integrate into their workflows.

Participant comments further reflected their excitement:

• “It’s like having a mind-reading coding buddy.”
• “I couldn’t imagine working without GitHub Copilot now.”
• “A magical experience—unless, of course, the AI suggestions were off-target!”

Clearly, the pilot outcomes were enough to pave the way for broader adoption.

Scaling GitHub Copilot across the enterprise

After the pilot’s success, it was time to move from testing to full implementation. Thomson Reuters created a four-stream strategy to scale adoption effectively:

1. Rollout and adoption: By automating license activation and usage tracking, they efficiently onboarded over 2,000 seats for developers across multiple teams.
2. Enablement and training: They established a community of champions—early adopters who became advocates for the tool—and offered in-depth training resources to empower them to make the most of GitHub Copilot.
3. Measurement and reporting: Metrics such as developer efficiency, code quality, and satisfaction, combined with monthly surveys, ensured that any challenges were addressed promptly. This approach also provided a clear way to track ongoing success.
4. Best practices and continuous improvement: The team shared success stories and use cases internally to maintain enthusiasm and optimize strategies for long-term ROI.

Sustained performance metrics

By expanding their data collection and feedback loops, Thomson Reuters observed the following:

• A 45% decrease in pull request (PR) durations.
• A 44% increase in the number of changes per PR, indicating more impactful contributions from developers.

Even more striking? Developers reported using their newfound efficiency to engage in learning, research, and innovative pursuits, enhancing their overall job satisfaction.

Want to ensure a successful GitHub Copilot rollout?

Make data-driven decisions by tracking active and engaged users, language preferences, and IDE usage. 

Explore the metrics guide now >

Lessons for enterprise leaders

For enterprise leaders like you looking to adopt AI tools like GitHub Copilot, here are the timeless lessons from Thomson Reuters:

1. Start small and scale

Pilot programs allow enterprises to test the waters, collect precise feedback, and refine adoption strategies before committing to full-scale implementation.

2. Focus on the metrics that matter

Establish quantitative and qualitative metrics early on (even before you start your trial). Whether it’s coding speed or developer satisfaction, identify KPIs that matter to your organization.

3. Invest in training and support

Don’t assume everyone will adapt naturally to AI tools. A structured enablement program ensures developers of all skill levels can benefit.

4. Build a community of champions

Identifying enthusiastic early adopters who can inspire others is crucial. Their success stories establish trust and excitement for broader teams.

5. Keep it collaborative

AI tools like GitHub Copilot thrive on collaboration. They’re designed to complement human effort—not replace it. Promote this mindset to avoid overreliance or misuse.

6. Iterate and adapt

AI adoption is a living process. Build feedback loops into your strategy to continually refine and improve as you learn more about the technology’s impact.

AI in enterprises is here to stay—are you in?

GitHub Copilot is a profound step toward the future of software development. By freeing up developers from repetitive coding tasks, it empowers them to focus on creative problem-solving, big-picture thinking, and innovation.

Thomson Reuters provides a clear roadmap for enterprises to successfully integrate AI, from thoughtful piloting to seamless scaling. Their outcome? Higher efficiency, better code, and happier developers.

If you’re ready to follow in their footsteps, get started now >

Thoughts? Questions? We’d love to hear your feedback and experiences with adopting AI tools like GitHub Copilot. Share your comments below. 👇