Working Together to Make Unit Testing Faster and Better: How BPM Company, Rabobank, and Pega Use Gen AI for Unit Testing

Why the Test Case Generator?

Within Rabobank’s testing strategy, unit testing plays a central role. This priority within the existing testing approach was the key reason why the unit test case generation use case became the first initiative to be developed in early 2026.

The challenge is familiar to almost every software development organization: writing high-quality unit tests takes time. Developers also tend to view it as one of the less appealing parts of their work, which can negatively impact both test coverage and quality. The project’s guiding principle can be summarized in a single motto: reduce the cost of quality. By making unit test generation less time-consuming and less labor-intensive, it becomes possible to consistently improve both the coverage and the quality of unit testing.

 

The solution: RuleIQ

Together, Rabobank, BPM Company, and Pega designed a solution called RuleIQ. RuleIQ supports developers in two ways:

1. Generating test scenarios – A Pega AI Agent gathers information about the relevant rule and the context in which it is used. This information is sent to a Large Language Model (Claude Sonnet), which generates proposed unit test scenarios. The developer reviews, refines if necessary, and approves the suggested scenarios.
2. Automatically generating the tests – Once the proposed scenarios have been approved, Pega automatically generates the corresponding Unit Test rules.

This approach demonstrates the flexibility of the Pega platform. Through AI Agents, Pega can integrate seamlessly with external Generative AI solutions, making applications like RuleIQ straightforward to implement. It is an architecture that can be leveraged for many use cases beyond this specific project.

 

From POC to enterprise-wide rollout

The project followed a structured implementation approach:

1. Proof of Concept – Validation of the technical concept and the feasibility of the proposed approach.
2. Development – Development of RuleIQ by a joint development team consisting of two BPM Company engineers and one Rabobank engineer, working closely with Pega. The project was managed by a BPM Company project manager.
3. Pilot – A number of existing Rabobank software development teams, known as pilot squads, will be the first to use RuleIQ and will conduct the beta testing of the application. The pilot phase will initially involve ten teams.
4. Rollout – Following a successful pilot, RuleIQ will be rolled out across all Pega development teams within the bank.

 

Business benefits for Rabobank

The expected value of RuleIQ for Rabobank is clear:

• Time savings, as RuleIQ generates test scenarios and automatically creates the corresponding tests once they have been approved.
• Higher-quality unit tests and improved test coverage, because creating comprehensive unit tests requires significantly less time and effort.
• More capacity for software development, as development teams become more productive by spending less time writing tests manually.
• Demonstrated flexibility of the Pega platform, showing how external Generative AI solutions can be easily integrated through Pega AI Agents.

 

Looking ahead

RuleIQ is the first of the ten Gen AI use cases Rabobank identified for its software development lifecycle. With the pilot serving as the next step toward an organization-wide rollout across 75 development teams, RuleIQ represents not only a successful first implementation but also the foundation for future Generative AI initiatives within Rabobank’s software development lifecycle.