Writing and Maintaining 2 Million Lines a Year Using Amazon Q Developer with BT Group

BT Group builds and runs some of the largest fixed and mobile networks in the United Kingdom. In addition to broadband, mobile, TV, networking, and IT services, BT Group also provides handsets, gaming, and insurance to help customers connect, communicate, share, be entertained, and do business.

BT Group uses AWS as its primary cloud environment, so when generative AI burst onto the scene, it implemented an AI coding assistant from AWS—Amazon CodeWhisperer (now part of Amazon Q Developer)—in September 2023 to accelerate code development. Then, when Amazon Q Developer became generally available in April 2024, BT Group quickly decided to adopt the tool. “Amazon Q Developer brought great new capabilities such as the in-console chat functionality that ticked the boxes on what we were looking for, so the pivot to that tool was almost automatic,” says Noppen.

A dedicated AWS team worked with BT Group to engage and train its developers in using Amazon Q Developer. BT Group focused on encouraging organic adoption across the company, offering training and licenses in particular to developers who contributed regularly to the company’s code repository. It also engaged with senior employees across business units to help them understand the benefits of the tool. The AWS team hosted hackathons, provided training materials, and helped BT Group identify champions of the solution to boost enthusiasm and facilitate learning across the company.

Because developers use Amazon Q Developer differently depending on the business unit, project, and coding language, AWS and BT Group tailored the various engagements rather than taking a one-size-fits-all approach. “It’s hard to overstate how important it has been for BT Group to work with the AWS adoption team to get people to use and understand Amazon Q Developer,” says Noppen. “The dedicated team helped us drive adoption to the level of use that we see today.”

Amazon Q Developer was fully implemented at BT Group by mid-November 2024, with around 2,000 active users. BT Group was eager to adopt Amazon Q Developer because of the new features and unique AI-powered agents that the service offered, including chat, code transformation, and testing capabilities. “After using Amazon CodeWhisperer for 9 months, our people were saying that it would be great if the tool could have chat functionality or focus on testing tasks,” says Noppen. “When Amazon Q Developer came along with these new capabilities, adopting it was a no-brainer.”

AI-powered chat from the console in Amazon Q Developer is a key feature that BT Group developers use to cut down on time spent searching the internet for information or best practices, thus increasing the time they can spend coding or building solutions. “The chat modality is the standout feature that has really changed and improved things for us,” says Noppen. “When the in-console chat feature arrived with Amazon Q Developer, we could see usage go up in the company, and we also saw higher acceptance rates.” Acceptance rates—how often a suggestion from the AI is accepted by a developer—doubled for BT Group for any interaction that included the chat feature.

BT Group developers also use Amazon Q Developer to generate unit tests. “Testing is important, but nobody likes writing test code,” says Noppen. “Having AI help out with this is very useful, and we’ve seen a sizable chunk of test code generated using Amazon Q Developer.”

Another useful feature of Amazon Q Developer is the in-line chat feature, a streamlined chat capability where developers can give feedback on the in-line code suggestions generated by the AI. BT Group also uses the Amazon Q Developer transformation capabilities, which can transform project code from older to newer versions that are better maintained and pose fewer security issues. This type of update, although important, often fell by the wayside previously because it took so much time.

Amazon Q Developer generates over 2 million lines of code per year for BT Group, compared with 2.5–3 million lines of code that developers previously produced annually. “Part of that previous 2.5 million lines of code is now being written by Amazon Q Developer, and Amazon Q Developer also adds some stuff over the top to help us solve other problems,” says Noppen. BT Group also uses the AI tool to maintain code that has already been written.

The solution streamlines processes and workflows for the company. Using AI, BT Group can accomplish more—not only writing more code but also testing, updating, and maintaining existing code—with the same amount of people. “Our jobs as we know them will change over time, and this is the golden opportunity for our employees to learn to use this tool well so that in the future they are more employable,” says Noppen.

Using Amazon Q Developer has simplified development for BT Group. Junior developers can take on difficult issues earlier in their careers because using Amazon Q Developer makes these cases more manageable. And because the solution streamlines workflows, developers can work faster and more efficiently using AI to get projects across the line.

“Because we have AI in our back pocket, what used to be a hard problem now becomes manageable,” says Noppen. “We can be more ambitious in the way we develop our software because Amazon Q Developer creates the operational bandwidth for us to do so.”