Test Guild's AI self-healing automatically fixes broken selectors when UI changes, but now flags every fix with email alerts and before-and-after screenshots. Developers see exactly what changed and can override the fix, file a bug, or accept it.
So if you were using Selenium, then at some point you have selectors somewhere, right? What we're doing is not black magic. We at some point need to interact with the device. And so in this case, we transformed your intent of enter this email into some selector, but that's not your problem. AI is doing it. The one thing we added since June, Joe, is that now when AI made a decision to change something technical, it will be flagged and the owner will get an alert in their inbox saying, hey, the app changed. I had to make a decision and I made this decision so you can audit it. Nice. And that closes the gap between AI and non-AI because when we show how AI is self-healing, one of the pushbacks you can get from developers is, hey, If I was using Selenium, it would have broken and I would have made the decision, is this a test issue or is this a product issue? If you auto-fix it, then you're depriving me from the opportunity to catch a bug. Yeah. So now we think this is the best of both worlds. We fix it for you, but it's auditable and we explain it to you. We say, hey, this is how it was last time I ran it. This is how it is now. And we also give you a screenshot just in case you, you wanna double-check the, the visual.
So, are you able to flag this as a proof? Like, okay, this is correct, or no, it's not correct, so it knows that any other test that runs it, that's not the correct result?
Oh, no, the idea behind it is that it's more— yes, at the end of the day, that's what we do, but here, what it is doing is that, using its intelligence, it is assuming it's correct, okay? It thinks it's correct because the intent is correct, it looks the same, but there was a technical change that probably you didn't wanna have, so it is going to consider the test as successful, but it will say it's successful, with something to look for, okay? And everywhere in the interface, if I go back to the history of runs, it'll have this icon that says, hey, it is successful, but I actually, it is only successful because I had to self-heal it. Gotcha. Along the way. And so you as a user, you can then override and say, I will file a bug for this and get it fixed.
Nice.
And so we stop depriving you from that opportunity to catch a bug. So we call that visible self-healing or auditable self-healing, and it's the number one requested feature from our customers to say, you know, I hate self-healing sometimes if you don't bubble it up like this.
So this is a great feature.
Exactly, yes. And it's everywhere now in the product. So if you go to analytics, then, you know, I can— this is my— this is the Test Guild demo, so it doesn't have a lot of runs, but I can go here and see that, hey, I had one test that self-healed, and I can, you know, I can drill down and go look at that.
