Major insurer

Mitigating the impact of change

To manage millions of client policies and products quickly and efficiently, the insurance company makes heavy use of automated straight-through processing (STP). Taking this approach puts a premium on software testing: the company must continually validate its systems to ensure that they are working reliably, accurately and efficiently.

If there are significant changes in the systems being tested, developers will need to make corresponding changes to existing test scripts. This was certainly the case for the insurer when it planned a major change to the kernel of its SAP ERP environment, which is at the heart of its STP capabilities. The change would impact many of the ternary identifiers of elements named in test scripts and could potentially cause them to fail.

To address this immediate requirement and simplify the ongoing maintenance of scripts for functional testing, the insurer looked for a way to mitigate the impact of changes on its existing testing automation solution, OpenText™ UFT One.

A spokesperson says: “We wanted to minimize the impact of the coming changes to our SAP environment, but we also aimed to increase the robustness and maintainability of our test scripts in general. We’re finding that it’s becoming harder to recruit highly experienced testers, so we looked for an approach that would empower new starters to create and maintain test scripts more easily.”

Now that the team has experience of the AI functionality in UFT One, we have cut scripting time from six hours per test case to just two hours.

Accelerating the creation of test scripts

OpenText™ (formerly Micro Focus), proposed that the insurance company use the new artificial intelligence (AI) capabilities embedded in UFT One to boost the resilience of its test scripts and increase the level of automation in regression testing. After rolling out the new functionality faster than expected, the insurer has already migrated ten applications and their 200+ test cases from property-based approaches to AI-based approaches in UFT One.

“We have a large and complex application landscape with 1,400+ automated test cases,” says the spokesperson. “These test cases run constantly: in 2022, we achieved approximately 27,000 test executions with UFT One. It took around four months to migrate the 200+ test cases to AI execution, and the set-up was slightly faster than the property-based set-up, even though our people had zero experience with the new functionality at that time.”

Because the AI-based execution of test cases can dynamically adapt scripts to cope with changed identifiers, future changes in SAP or other third-party applications will not negatively impact testing. The solution also accelerates the creation of new test scripts. The spokesperson says: “Now that the team has experience with the AI functionality in UFT One, we have cut scripting time from six hours per test case to just two hours.”

The insurer uses OpenText™ ALM Octane to manage the test cases for UFT One, and OpenText™ LoadRunner enterprise to manage performance and load testing.

Using the AI capabilities in UFT One has accelerated our test scripting, reduced the maintenance significantly, and made us far more resilient to changes in applications.

Reducing maintenance while boosting resilience

So far, the insurer has cut 60 percent from the time required to script a new test case, and the adoption of AI in UFT One has dramatically reduced the maintenance of test cases, too. The company has completed more than 1,000 execution cycles with zero need for technical maintenance or changes to the AI-created scripts. This long-term resilience reduces effort, freeing up skilled staff to focus on other priorities.

Getting new users up to speed is much faster now: people without previous experience in creating AI-based UFT One scripts require only two hours of training, versus around seven hours for property-based scripts. “It’s not just that we can train people faster, but also that the resulting scripts are of higher quality than before” says the spokesperson. “This is because people can focus on the testing logic rather than needing to worry about recognizing elements or understanding how the properties are stored.”

The spokesperson concludes: “Using the AI capabilities in UFT One has accelerated our test scripting, reduced the maintenance significantly, and made us far more resilient to changes in applications. Overall, this strengthens our ability to ensure that applications deliver the core functionality as expected and that we minimize incidents in production.”