The challenge
A global energy major was standing up a greenfield Endur implementation. Greenfield builds are first-principles territory: there's no production baseline to compare against, so every test has to assert what the right answer should be from the specification, not from yesterday's behavior.
Combined with a Scrum delivery cadence, that puts pressure on the testing function from the first sprint. Defects that leak from in-sprint into integration, regression or production carry compounding cost the further they travel. The question wasn't whether to automate testing. It was how to embed it deeply enough, early enough, to keep up with the build.
The solution
The shift-left approach put Triangle into every Scrum team. Test automation engineers embedded with developers and business analysts, writing in-sprint tests against new functionality as it was built. Overnight regression ran against the cumulative platform, catching anything that had drifted between sprints.
Triangle's integration with Kafka extended coverage into the messaging layer, confirming that trades, prices and downstream events flowed correctly between Endur and the wider trading architecture. The automation estate grew alongside the build, rather than being retrofitted at the end.
The results
- 265 defects surfaced through Triangle in-sprint tests and overnight regression, caught before any of them could reach production.
- Around 60,000 automated test step executions ran overnight, covering the platform as it matured.
- Triangle was embedded in each Scrum team, making test automation a first-class delivery artefact rather than an end-of-cycle bottleneck.
- Kafka integration extended assurance to the messaging layer, where end-to-end behavior is hardest to verify by hand.
- Test assets built sprint by sprint formed the basis of the platform's living documentation, ready for ongoing regression and future upgrade cycles.
Why Trinitatum?
Greenfield isn't regression. There's no baseline yet, so first-principles is the only viable approach. We bring the technical judgement to know which testing style fits which situation, and the depth in Endur to make it work alongside Scrum teams from sprint one.
