image

Trends

TTC at Tricentis Virtual Summit: Machine learning in software testing

By Daisy Wang, Marketing Associate at TTC

With the increasing prevalence of AI and machine learning, there’s no doubting the role they’ve played in sparking an influx of new tools within the software industry. However, as with any new wave of transformation and invention, it’s necessary to evaluate where the real value and ROI lies. At this year’s Tricentis Virtual Summit, TTC Senior Manager Nate Custer delved into which organizations can truly benefit from utilizing machine learning/AI while separating fact vs. fiction when it comes to these trends.

Through an analysis of three specific use cases of machine learning within testing, Nate breaks things down beyond blanket statements and overreaching promises to paint a realistic view of where TTC is seeing ROI for clients right now. Here in particular, he shines a light on autonomous testing, self-healing selectors, and test selection. Or – as Nate puts it – the strawman, the silver bullet, and real-world ROI.

When conversations around autonomous testing pop up, it seems unavoidable for questions along the lines of, “Will AI replace human testing?” to make an appearance too. In short: no, it won’t. As Nate covers within his session, even the best applications are built and refined by human testers because no algorithm will behave flawlessly on its own. Issues with software models are inevitable, and it’s up to real people to identify them and deduce where effective solutions can be found.

While machine learning is a powerful new technique to solve problems, it’s not always the best solution for an organization. For example, when it comes to self-healing selectors, those building automation for the first time or lacking familiarity with industry-leading tools would make for an ideal candidate. On the other hand, those who have applications built on widely used commercial platforms may be better off opting to continue using their existing platform. According to what Nate has observed so far, self-healing selectors don’t reduce maintenance enough to be worth replacing existing automation.

Finally, Nate shares TTC’s results with ML tools for test selection/impact analysis. When it comes down to it, there’s no definitive one-size-fits-all solution. Evaluating pain points, organizational needs, and individual context are key to understanding if these new tools are the right fit.

If you missed the Summit and are curious about how ML/AI could help your organization, an on-demand version of Nate’s breakout session, Machine learning in testing: Cut through the hype and achieve real ROI, is now available to stream at any time. From its inception, TTC has been passionate about providing impactful testing solutions and engaging with our clients about how our services can best suit their needs. To learn more about TTC, visit our website.