How test automation can smoothen the delivery of new products and systems

As products and systems are being introduced and upgraded at an ever-quicker pace, the importance of thoroughly testing the technological changes before go-live is increasingly at the forefront of IT agendas. With testing for long has been notorious for its cumbersome work and high-workload, smart technology is now coming a long way in making the whole process smoother and more efficient.

The testing process of any new system or product can be a long and arduous journey. It involves repetitive but necessary processes, such as testing processes, dependencies, performance, and more.

This form of extensive testing is essential in modern software development, says Charlene Venter, Consultant at Varrlyn. “It safeguards the quality of the delivery, and ensures that new systems or products meet the requirements and demands set by the business and IT teams for coding, functionality (business logic), and user experience.”

“The risks associated with not uncovering issues prior to launch can be huge, from loss of customer satisfaction, to reputation damage, financial damage, or outage of other applications and systems, to name a few.”

With technology making more and more possible, the domain of testing is emerging as a large beneficiary. New tools are enabling fully automated testing, from the development of the test cases through to the actual review and validation of tests.

“Test automation can save huge amount of work,” says Philip van den Emster, Director at Varrlyn. “Not just that: it can enable the testing of much more scenarios and cases, enhancing the accuracy and overall quality of the process. It also ensures consistent test execution across different software development cycles.”

In practice

Headquartered in Amsterdam, the Netherlands, Varrlyn supports numerous financial services institutions with launching new products and/or upgrading systems. To that end, the consulting firm has embraced test automation in its approach.

Pointing to one example, Van den Emster highlighted a complex data hub project for a large institution. “Our client is setting up a centralised hub for all its portfolio data. This hub will serve as ‘single source of truth’, saving huge amounts of time in the long-term.”

But before the hub could go operational, over 1 billion of rows of data needed to be tested. “That was obviously not possible manually,” he says. “Instead, we came up with the idea of creating an automation test framework. This could take care of the core functionalities and GUI, which can take care of the data visualisation. And this also helps to maintain the test scenarios as part of the regression testing suite as multiple capabilities have to be re-tested when a new dataset is onboarded.”

Test automation can also provide significant value in tackling the large number of dependencies between processes and systems, especially in financial services, where institutions have a very complex integrated environment. “Indeed, they have financial applications from a lot of different vendors, internal and external, and these all need to work seamlessly together.”

“Trading platforms, market data feeds, back-office systems. All of these need to be integrated so that they can exchange data accurately and securely between systems,” Van den Emster continues.

Then comes the aspect of regulatory compliance, which adds a further layer to data complexity – and the need for robust testing. “This means there is a high level of data variability to contend with. And doing so can make all the difference when dealing with a constantly shifting regulatory environment.”

Interconnected tests

Illustrating all these points in a single example, Venter pointed to a particular case, in which Varrlyn supported a client with the introduction of a new back-office platform, while at the same time decommissioning the legacy one. As the new way of working and platform was notably more elaborate, the project also had a large upgrade element to it, such as introducing new functionality and upskilling workers.

During the project, test automation ensured that interdependencies were extensively tested for, and that different sets of tests were for both system transitions. Meanwhile, all regulatory requirements were included in the test cases, and elaborately checked. “Key here was ensuring interconnected tests between the different pillars,” says Venter.

Lessons leaned

Based on Varrlyn’s extensive experience with guiding the software testing process, Van den Emster and Venter said that the firm has established several best practices in the domain.

One of those in ensuring that testing should not be an isolated workstream, but integrated into the activities of the business and IT. That process should ideally start early on in the process, says Venter. “You need to organize the test function up front, involved with the team throughout all cycles.”

Agreeing to this, Van den Emster cautions against siloing the function of test automation to just the IT or test team. It should instead be “embedded in the business area using a certain platform” – as test automation is only useful as long as it “reflects the actual business it is simulating”.

Venter meanwhile adds that companies should focus on building the minimum amount of test cases with maximum coverage. “Keeping things simple and failing fast is advisable, as it enables an institution to adapt weaknesses quickly, and pivot how they are testing. At the same time, they should not aim for a 100% pass rate – the point of this kind of testing is to find failures, and it is a good thing when they are identified.”

Van den Emster notes that automated testing should be run “as frequently as possible”, as this will help catch any issues “as early as possible”. But reiterating the need to keep things simple, he adds that this only works “when the test suite is concise.”

The potential of AI

Generative AI is already being used to transform a range of business functions across the business gamut. According to Van den Emster, Generative AI can prove a game changer for making test automation smarter, more integrated, and accessible. AI and machine learning will in turn increasingly power the enhancement of test creation and execution.

“AI technologies can identify patterns, predict potential failures, and create more efficient test cases,” Van den Emster says. “AI will also assist in maintaining and updating test scripts automatically as the application evolves, reducing the need for manual intervention. Self-healing systems will even be used to automatically detect and adapt to changes in the application interface.”

“All of this will enhance the efficiency, coverage, and reliability of testing processes, ensuring that applications meet the high standards required in today’s fast-paced and complex technological landscape.”