I wanted to share a short video where I walk through our new Automated Testing capabilities in Infinity ‘25 and start a conversation with you on the areas where AI driven test automation can make a real difference.
Like many of you, I’ve seen teams struggle to achieve full test coverage. And it’s not because it isn’t important, or because people don’t want it - it boils down to the fact that writing and maintaining tests is time consuming and often pushed aside in favor of delivering new functionality.
In my video I show how Autopilot uses AI to generate testing scenarios and Gherkin scripts based on your application and case context, supporting both Case level and Application level testing. I also demonstrate our support for both automated API testing and UI testing (with playwright scripts).
Our goal with these capabilities is simple: make automated testing easier to include in everyday sprint work, so teams can move faster, deliver higher quality, and reduce technical debt over time .
You touched on this in the video: test creation doesn’t feel like value creation in the way that delivering a story or a bug fix does. From my experience, I’d add that this isn’t just a matter of how the builders feel about it, but all stakeholders. Getting a new test suite feels like opening up a birthday present and it’s… a pack of socks. “OK, yes, I’m down to like two pairs and one of them has holes, I should be thankful, and yet I can’t shake the feeling that this is displacing a cooler gift.” Taking the friction out of tests means I get my socks with the knowledge that they’re not taking the place of the cool things I actually wanted.
What we’ve seen historically is that even with 100% test coverage, the tests can be effectively meaningless because they’re not testing the intent of the application. Tests that exist soley for the purpose that “they clear a guardrail warning” have provided cold comfort, at best.
Automation can certainly help, but test cases and scenarios must first be prioritized, then automated, according to business need and risk. Test design needs to be baked in from the outset of the application build, even during Blueprinting. In short, tests need to be designed along with your features, but perhaps not with so much detail you are maintaining two sets of logic, as Anthony points out in the video.
All of that said, Autopilot for Testing is a huge leap forward. Human in the loop, always, especially when it comes to testing!
By adopting this feature, we move toward a shift-left quality model, allowing us to bake quality into the development lifecycle rather than treating it as an afterthought. Exciting times!
Watch the video here: