If you’ve flown on a plane or navigated by phone, chances are you’ve relied on Indra’s technologies to get you to your destination safely—after all, they are a leading provider of air traffic control systems and build technologies like the Galileo global navigation satellite system.  But that’s just the beginning. From aerospace to mobility to defense, Indra builds innovative technological solutions that incorporate AI, big data, and virtual reality into a new generation of critical systems that transform the industries and key operations of its clients worldwide. 

When the pandemic spiked demand for top talent in the tech sector, Indra turned its innovative vision inward to increase its developers’ productivity. It wanted to eliminate repetitive work while also making the company more appealing to prospective hires, so it built a dedicated team to evaluate different AI code generation tools, such as TabNine, Kite, Amazon CodeWhisperer, ChatGPT, and GitHub Copilot.

GitHub Copilot stood out from the start. After using it for just a couple of days, we saw that it was already giving us more appropriate suggestions. We knew it was just the beginning.

“GitHub Copilot stood out from the start,” says Ana López, Head of the Aerospace and Defense Software Development Center Unit. “After using it for just a couple of days, we saw that it was already giving us more appropriate suggestions. We knew it was just the beginning.” 

Indra’s first steps were to determine that GitHub Copilot would meet the strict privacy and security regulations of the industries it operates in. As detailed in the GitHub Copilot Trust Center, the Product Specific Terms laid out how their data would remain completely private. For example, when GitHub Copilot uses Indra’s codebases to provide context for code suggestions, that data remains encrypted and is not stored. 

The team decided to adopt GitHub Copilot on a non-confidential Java codebase to determine whether it could boost developer productivity and happiness, while also improving code quality. 

"With very little experience in the language I was coding in, GitHub Copilot helped me generate a fully-functional class in less than three minutes,” says Software Engineer Luis Miguel Guerrero. “It was truly awe-inspiring to watch Copilot generate 70 lines of code."

GitHub Copilot’s effects on productivity were also immediately apparent. 

“Copilot’s ability to generate unit tests automatically is truly remarkable,” says Software Architect Jorge Yago Fernández. “Writing comprehensive unit tests can be time-consuming and tedious, but with Copilot's assistance, we witnessed firsthand how it could rapidly generate test cases, saving us valuable time and effort, while also ensuring thorough test coverage.”

For a company like Indra, whose tools aim to make the world a safer place, there’s no room for error. Code quality cannot be sacrificed for velocity when you’re working on projects such as Galileo, the most precise satellite navigation system in the world. GitHub Copilot allows them to accelerate software delivery while meeting the same high standards.

“Copilot helps our developers write code faster, with fewer errors, and deliver software solutions more quickly,” says Joaquín Anaya, Senior Manager of the Aerospace and Defense Software Development Center Unit.

After that initial evaluation, Indra’s Aerospace and Defense Software Unit rolled out GitHub Copilot across a range of non-confidential codebases. Many of Indra’s developers were excited to get started, knowing that GitHub Copilot would help them with repetitive tasks and even with migrating to other technologies. The unit offered pair programming sessions and training with GitHub Experts to demonstrate its capabilities and bring everyone on board.

Soon enough, GitHub Copilot was an integral part of their workflow.

GitHub Copilot allows us to approach our daily work with such greater efficiency and agility.

"GitHub Copilot allows us to approach our daily work with such greater efficiency and agility,” says López. “When our team can't use it, we feel like we’re missing an essential tool to do our jobs."

With GitHub Copilot automating tasks like generating documentation, designing modules, and writing boilerplate code, Indra has seen a 20% boost in productivity when developing new features. Overall, the teams using GitHub Copilot have seen a 15% increase in the number of tasks in each three-week sprint the team completes, meaning they are delivering value at a higher speed.

“GitHub Copilot has been instrumental in reducing the cognitive load on our developers,” says Anaya. “It freed up 20% of our team's time and mental energy to focus on more complex and high-value aspects of development. It’s more rewarding work, which has led to increased job satisfaction and motivation.”

The newfound velocity and ability to handle complexity couldn’t come soon enough. Indra has also been making the move from waterfall development to agile, in an effort to take a more collaborative, iterative, and customer-centric approach to software development. As part of that, the company is redesigning some of its systems with a microservices architecture on Kubernetes. Much like how it helped Guerrero write a class in an unfamiliar language,  López says she expects that they will use GitHub Copilot exhaustively, as new programming languages and frameworks become part of the technological baseline.

Looking to the future, Indra aims to roll out GitHub Copilot across all of its non-confidential projects. So far, López says its effects have been transformative.

“GitHub Copilot enables us to work more efficiently, produce higher-quality code, and deliver projects in a timely manner,” says López. “It has empowered our team and enhanced their productivity, ultimately driving the success of our organization.