Another day, and yet another artificial intelligence-focused startup is launching itself out of the gate. The new company is called Harness, and its goal is to enable continuous software delivery for companies of all shapes and sizes.

For the uninitiated, continuous software delivery is a relatively new software development method that sees updates rolled out around the clock, often several times in one day. It’s a stark contrast to older software development techniques, where updates are often spaced several months apart.

Not surprisingly, continuous software delivery is hugely popular among larger enterprises that can benefit from rolling out changes to their critical business applications at will, but smaller and medium sized organizations are finding it much more difficult to get up to speed. The problem with continuous software delivery is that it’s very difficult to implement correctly unless you have the right skills and systems in place. Most smaller firms don’t, and so they struggle to deliver software updates quickly without errors, as they often run into problems that can only be fixed manually, much like traditional software development methods.

Harness is on a mission to change that reality, with the pedigree that gives it a chance to make it happen. The startup, which exited stealth mode today with $20 million in Series A funding from Menlo Ventures and BIG Labs, is led by Chief Executive Officer Jyoti Bansal (pictured), who previously served as CEO of application monitoring firm AppDynamics Inc. before it was sold to Cisco Systems Inc. for $3.7 billion in January. The company’s other co-founder is one Rishi Singh, a former DevOps platform architect with Apple Inc. who now serves as Harness’s chief technology officer.

“At AppDynamics, our customers were happily using our platform to monitor their complex software applications, but almost all of them told me that the process for delivering rapid changes to those applications remained a huge problem,” Bansal said. “Software engineering teams need a platform that’s intuitive and powered by modern AI to meet demand for incredibly fast, high-quality releases.”

That’s what Harness intends to deliver with its new AI-powered platform. Harness does away with the need for teams of skilled coders to monitor new deployments because it uses unsupervised machine learning technology to do much of the legwork instead.

With machine learning, Harness is able to understand what’s happening with an application’s baseline environment and initiate automatic rollbacks to a previous version if irregular activity is detected. This means companies can push ahead with new updates in the knowledge that they won’t suffer any downtime or failures. The platform also comes with a pipeline builder and workflow tools for developers to create updates for their apps before the system checks them for errors.

“The machine learning aspect of Harness give us the peace of mind and confidence to conduct multiple deployments on any given day,” said Ed Rose, director of software development at Build.com, which has been testing the Harness platform to implement software updates. “In a pre-Harness world we would spend maybe an hour with several of our senior engineers monitoring activity post-release. We can now reduce that to one individual who oversees Harness as needed, freeing up our engineering talent.”

Harness’ new platform is available now, and the company is hoping to get the ball rolling by offering free trials of its software.

Photo: Harness