Jetify, the company formerly known as Jetpack.io, is launching its first AI agent product Tuesday. Dubbed Testpilot, the company’s first AI agent is meant to make the majority of the routine test creation process a matter of simply letting Jetify create a test plan, execute it, and report back, all while the agent directly interacts with the user interface.
Right now, Testpilot works for web-based applications, but the team is also planning to launch mobile testing soon.
The company’s original focus was mostly on creating dev environments in the cloud. While it will still offer those (though maybe with a stronger focus on making those API-driven), Jetify CEO and co-founder Daniel Loreto describes Tuesday’s launch as a bit of a pivot and as the first in a series of AI agent launches, all of which will focus on improving the software development lifecycle.
“We feel, and I’m sure a lot of people feel this way, that AI is truly one of those once-every-30-year technology waves that comes along. It’s like the invention of the Internet or mobile, right? So having the opportunity to be part of that wave and to really participate in it convinced us that, yes, let’s take the risk. Let’s shift,” he told me.
The team decided that the best way to get started with AI agents was to work on something that solved a clear pain point for developers. Nobody likes writing tests — and when it comes to UI testing, existing tools like Selenium can quickly become brittle with even the slightest UI changes. That’s also exactly what AI tools are good at handling: They won’t break, even when a button moves a few pixels to the left.
Loreto also noted that QA teams are already used to using third-party tools for testing their applications, making Testpilot an easier sell. “It’s a well-defined motion that companies say, ‘Yeah, we need this end-to-end testing, and we’re willing to hire a third party to help us with that,’” he said.
He also stressed that this kind of end-to-end testing is easy to integrate into the development process since all the company needs from the developer is access to log in to the application that the service is meant to test. Since Testpilot looks at the actual application, not the underlying code, it doesn’t need access to GitHub or any other code repository, for example.
However, even though Testpilot today can cover 60% to 70% of tests that humans would likely come up with, there is still a gap. That’s something Jetify acknowledges, but as Loreto noted, he thinks that this is still “a huge value-add.”
As of now, potential users still have to go through an onboarding flow with Jetify. Soon, the team hopes to launch a self-service version as well (with mobile support following after that). Jetify is still experimenting with how best to price the service, which is, in part, why the company hasn’t launched a self-service version yet.
