A bank’s software developers can’t get by on speed alone.
Deloitte IndustryAdvantage and AI Assist helped a bank transform its software development life cycle so teams could work smarter (not just faster), together.
THE SITUATION
A global bank wanted its software developers to write high-quality code faster. Like all major banks, it faced continuous demand for new customized software on ever-accelerated release cycles. And the stakes seemed higher than in many other industries—in digital banking, downtime faces serious scrutiny; bugs can make headlines.
But no matter how quickly developers may write code, no digital banking offering hits the market from a lone genius hammering away at a keyboard in a closed room. Each new offering and update takes a web of product owners, architects, developers, and quality assurance engineers to guide the software development life cycle (SDLC) from requirements gathering all the way to rigorous user acceptance testing. Really, the bank didn’t need its software developers just writing code faster. It needed its whole digital banking team working smarter, together.
Across each digital banking department—and from team to team within the same department—different technology tools, subvariants of tools, SDLC processes, and subprocesses abounded. Add on a large network of external software suppliers, and the bank lacked any kind of common dialog across the SDLC. Because of this, teams spent significant time and energy translating between different working styles, scheduling meetings, volleying questions back and forth, whiteboarding, drawing diagrams, double checking on this user story or that requirements framework.
At the same time the bank’s line of business (LOB) owners were considering these challenges, buzz around Generative AI (GenAI) was growing and crowding in on all sides. The LOB owners each saw clear potential, but a less clear path to harnessing that potential. They decided to put new GenAI-assisted SDLC processes to the test in their own environments to find out if their teams could fundamentally transform how they delivered software.
IT’S A TALE AS OLD AS TIME. OR AT LEAST AS OLD AS DIGITAL BANKING.
THE SOLVE
The bank began the search for GenAI solutions to integrate into various points of the SDLC. But through conversation with Deloitte, digital banking leadership realized that rather than just using point solutions to replace isolated tasks, they could derive much greater value from a true end-to-end GenAI solution that could unify team members across the enterprise's entire SDLC. They found that end-to-end tool in Deloitte’s proprietary AI Assist™, which embeds GenAI capabilities into existing workflows to create a more cohesive process, build better-quality code, and achieve quicker software delivery. But beyond the tool itself, the Deloitte IndustryAdvantage™ framework—characterized by deep banking sector experience and a problem-solving approach—could help cut through all the noise around GenAI and achieve the new ways of working and common dialog the bank needed.
Together, the bank and the Deloitte team envisioned a proof of concept: deploy AI Assist to one LOB for eight weeks in a controlled environment with a select group of about 15 participants from the client’s and Deloitte’s product owners, engineers, and quality assurance teams. Through the proof of concept they would identify where they could introduce new ways of working and how they could customize AI Assist to accelerate those new processes and communication norms.
Deloitte specialists examined each role along the software development journey and reimagined how each person could more effectively hand off the work to the next person, instead of getting snared in the usual web of meetings, follow-up meetings, and return for rework messages. They began helping bank product owners use AI Assist to transform business requirements into user stories that drew on prior effective user stories as context, replicating the stories’ strengths. They then passed these more familiar, context-informed stories along to architects and developers who used AI Asist to automate the mundane, time-consuming parts of creating flow diagrams or generating unit tests—freeing up more time to focus on creative problem-solving and integration across the wider LOB or enterprise tech stack. QA engineers could then achieve even higher test coverage from AI Assist’s autogenerated integration tests.
Over the eight weeks the bank's team started to notice less time in meetings, more cohesive collaboration, and an all-around new sense of harmony across the SDLC. As team members relayed tips about effective prompts and experiences on Agile retrospectives, they learned to use the same language, created new shared norms, and began to glimpse the potential of a powerful AI tool combined with strategically designed SDLC processes.
FROM DEVELOPERS’ FINGERS FLYING TO THE WHOLE MACHINE HUMMING
THE IMPACT
The transformed, GenAI-assisted ways of working revealed significant potential gains:
Improved productivity.
Over the course of the proof of concept, the team saw productivity improve 53% across the SDLC tasks tested. And when the more seasoned team members paired AI Assist with their knowledge of underlying systems and processes, the productivity gains were even stronger, demonstrating the power of technology combined with human insight.
More consistent delivery.
AI Assist helped the initial pilot team reduce time spent on repetitive tasks like updating documentation or creating test cases, speeding up their potential time to market for new software and potentially giving the bank's customers quicker access to new features to fit their evolving digital banking needs. By increasing test coverage, the solution also holds the potential to decrease defect density and unexpected bugs—especially important given the urgency of digital banking.
Enhanced talent experience.
Empowered by leading-edge technology, team members could focus more time and effort on rewarding, strategic, and value-added work, improving their talent experience.
Word of the results achieved in the proof of concept reached the bank's other LOB owners and inspired them to consider a similar transformation for their units. The bank continues to explore full-scale adoption of the new ways of working and the potential to integrate AI Asist into its enterprise AI stack, collaborating with Deloitte to understand the nuances and change management implications of an enterprise-level rollout, including careful consideration of how the appropriate productivity measures may vary across teams. This continued exploration opens up more possibilities for the bank to unite around a common dialog and powerful new ways of working.
INSTEAD OF SPEED ALONE, TRY PEOPLE, PROCESS, AND TECHNOLOGY TOGETHER.
Ryan Lockard
