GitHub Copilot at Bankdata: A Great Developer Experience Increases Our Ability to Build and Transform Applications

With the implementation of GitHub and most recently GitHub Copilot, Bankdata has elevated its code developing skills to a new level. The AI tools have significantly optimized workflows and increase both job satisfaction and efficiency in software development.

In collaboration with Microsoft, Bankdata ran a pilot project initially with GitHub and later with Copilot for GitHub as part of a long-term strategic effort to continuously evolve as a modern tech company in the financial sector. Bankdata is one of Denmark’s largest software companies and provides technology solutions to the financial sector. Implementing AI tools in a company that operates in a compliance-heavy sector can seem like an overwhelming process, but despite the resources spent on compliance considerations and revision of internal policies, the benefits of the tools have made the effort worthwhile for Bankdata.

The cloud-based platform for developers – GitHub – has already provided a more enjoyable and efficient way for developers to work and collaborate. Additionally, as Bankdata continues to consolidate source management around Github, the need to manage on-prem solutions is eliminated. Finally, GitHub Copilot has significantly improved the company’s developer tooling, increased market competitiveness, and improved the ability to attract and retain talent, explains Kenneth Fiil, CTO at Bankdata.

What we have achieved with GitHub Copilot, I find hard to see that we could achieve with any other product in the market right now

Kenneth Fiil, CTO at Bankdata

Improved Developer Experience and Higher Quality Code

It has been crucial for Bankdata to experiment with GitHub Copilot among developers on both COBOL, a 65-year-old programming language for mainframe computers, and the more modern and widely used Java language, to ensure it can create value for as many as possible.

A proof-of-value study showed that the value far exceeded expectations, both in terms of efficiency and value creation, but also for the individual developer’s daily work, explains Michael Lind Mortensen, Head of Source Management at Bankdata.

“86 percent of our developers improved their developer experience with GitHub Copilot. They spent significantly less time troubleshooting and could thus maintain their flow and ultimately deliver higher code quality.”

Michael Lind Mortensen thinks that GitHub Copilot has been particularly interesting in relation to Bankdata’s mainframe technology, which drives some of the most essential products they offer their customers.

We have had the opportunity to stress-test how GitHub Copilot handles our COBOL layer specifically, so we can continue working with the existing portfolio while transforming it

Michael Lind Mortensen, Head of Source Code Management, Bankdata

Can Change Legacy Challenges in the Long Run

It is important for Bankdata that their developers work with the best tools in the industry and can focus on complex issues without dealing with complex infrastructure, thus creating value with as little friction as possible. But Kenneth Fiil not only sees the potential for creating value in existing processes. It also applies to the ability to attract the industry’s best developers.

To create the best environment to do so, Bankdata has had a broad compliance focus to ensure that it is also possible to bring other products or AI tools into the
business in the future.

It’s actually not just about the risk of introducing a new tool, but perhaps even more about the risk of not doing so

Kenneth Fiil, CTO at Bankdata

Because by offering the opportunity to train and develop skills on the latest generative AI tools, Bankdata becomes a more attractive workplace.

In the long run, Kenneth Fiil sees that GitHub Copilot and the tools that will follow in its wake can not only increase efficiency and product development but also change the way companies like Bankdata address legacy.