For professional services organizations like Accenture, staying on the cutting edge of technology adoption isn’t an option, it’s a requirement. More than 9,000 clients in 120 countries rely on Accenture to provide them with the most innovative technology and consulting services, helping its clients adopt everything from generative AI and blockchain to edge and quantum computing. With such a comprehensive portfolio of services, Accenture has turned to GitHub Copilot Business, adopting it on an enterprise scale to not only boost productivity and developer satisfaction but also to enhance the value delivered to their clients.

After an initial pilot with just 20 developers, Accenture partnered with Microsoft to conduct an extensive, randomized controlled trial (RCT) to measure Copilot’s enterprise impact. Measuring the performance of 450 developers using Copilot, compared to a control group of 200 who were not. The results were convincing. Accenture found that:

• We saw a significant increase in successful builds, enabling developers to spend less time debugging and to deliver projects faster.

• Developers generate more pull requests per week, accelerating the overall development process and shortening the time to market for new software releases or updates.

• Accenture's Developer Experience improved, with 95% of developers reporting they enjoy coding more when using Copilot.

They have since rolled out Copilot Business to more than 12,000 developers.

Writing tests is one realm where Accenture’s developers have found Copilot to be extremely useful. “It’s allowed us to take the time to create all the unit tests, functional tests, and performance tests that we want in our pipelines without having to go back and effectively write double the code. There's never been enough time in the past to go back and get to all of them,” said Schocke.

In addition to writing tests, Copilot has also allowed Accenture’s developers to tackle the ever-increasing technical debt that challenges any organization of its size.

“We have more work than developers. We just can’t get to it all,” said Schocke. “By augmenting our developers’ skills and helping them to produce features and functions more quickly with higher quality, we’re able to get to more of the work that just didn't happen before.”

One way developers use Copilot to handle technical debt is by asking Copilot Chat to explain legacy code. Then they can feed that explanation, along with additional parameters and updated information about the environment, back into Copilot to help refactor and modernize the code.

We want to do the best for our clients that we possibly can. GitHub Copilot enables us to do that because it allows us to move faster and developers to come up to speed more quickly to work in unfamiliar areas with more confidence.

“We want to do the best for our clients that we possibly can,” says Kristine Steinman, senior manager of Generative AI at Accenture. “GitHub Copilot enables us to do that because it allows us to move faster and developers to come up to speed more quickly to work in unfamiliar areas with more confidence.”

“It's easier to share code across project teams with GitHub, which was designed from the ground up to facilitate collaboration,” said Schocke. “We're trying to get out of the business of duplicating efforts, and now we can do a lot more internal code sharing without having to worry about managing additional permissions.”

With GitHub Enterprise, Accenture’s developers can innersource common solutions that can be customized for clients instead of reinventing the wheel every time they get a similar client request. To promote this use case alongside Azure DevOps, the company has configured access to internal GitHub repositories using Microsoft Entra ID (formerly Azure Active Directory). 

But integration doesn’t end there. Accenture also manages all of its enterprise users with IssueOps, triggering automations with GitHub Actions for tasks like creating repositories, adding team members to a repository, and other administrative tasks. By consolidating and automating these tasks, time once spent managing multiple tools can now be spent on more valuable activities.

Finally, Accenture’s Information Security team now mandates that all GitHub users adopt GitHub Advanced Security, to ensure no secrets or credentials are leaked across their innersourcing work. While the company used to build features to handle this themselves, Schocke said he’s happy to take advantage of GitHub’s features like Dependabot and Secret Scanning.

“We realized that building and managing our own tools costs us more in the end,” said Schocke. “Using GitHub Advanced Security has changed the mental model for our security leadership. Before, we relied on somebody manually adding that scan to the build or deployment pipeline. Now, every single bit of code that gets committed to our repository gets scanned with Advanced Security.”

Going from a pilot of just 20 developers testing Copilot to a roll-out to 12,000, Accenture has not only accelerated productivity and increased satisfaction within its developer organization, but also realized the benefits of GitHub’s AI-powered developer platform.

GitHub helps us build better solutions, faster. And in the end, it’s our clients who benefit.

"GitHub helps us build better solutions, faster," said Schocke. “And in the end, it’s our clients who benefit.”