Challenge
Manual BDD processes leads to inaccuracies and complexities
BDD tests are designed to streamline the software development life cycle. Using a BDD approach, developers can build and test apps and programs based on the behavior a user expects to experience.
However, developers at this investment management firm were manually creating BDD scenarios. Human error and bias led to incomplete, unreliable, and inaccurate testing scenarios. As the complexity of each application grew, the time and effort required for testing also grew, burdening the developers.
With manual BDD testing, it was also difficult for developers to adapt and respond to evolving business requests – limiting their agility and responsiveness.
To further complicate things, the testing life cycle and test generation process lacked traceability. Without a clear process, employees struggled to map test scenarios against user and business requests.
Solution
Redesigning processes with generative AI
Genpact stepped in to redesign the BDD testing process. We knew the goal was to introduce gen AI into BDD testing, but first, we needed to improve the company's infrastructure.
In just three months, we transformed the company's infrastructure using Flask – a micro web framework written in Python and hosted on a private cloud. We integrated this with internal project tracking software and large language models. This integration meant BDD tests could be directly mapped to user requests.
We effectively automated the creation of test cases based on desired application behavior to improve testing efficiency, traceability, and reliability.
We then introduced gen AI. By analyzing user experiences and specifications, AI generates realistic test scenarios mimicking real-world interactions – drastically reducing the manual effort required.
Impact
Traceable, reliable, productive
Creating this modernized infrastructure and process for BDD testing revitalized the company's software testing life cycle. With gen AI at the heart, the company achieved:
Faster turnaround: Automated BDD scenario creation and mapping freed up one developer per day to focus on high priority tasks
Efficient software development: Teams can now move through software development cycles in half the time
Versatility: Teams can conduct comprehensive testing across a broader range of scenarios
Traceability and visibility: For quality assurance teams, business analysts, and developers
Scalability: AI-driven BDD allows teams to handle larger and more complex projects with ease
Reduced time to market: Automated scenario generation and validation accelerates the software development life cycle of new features and updates
Now, developers spend less time crafting test scenarios and instead focus on reviewing the scenarios created by generative AI. Plus, they can focus their attention on higher-value work. The company also benefits from a solid AI foundation to support future software projects and continued business growth.
Related case studies
- one
- one
- one