AI-assisted software development saves time and reduces errors, according to one company that has introduced a low-code visual platform for building customer enterprise software. “Something that took [developers] … five minutes, now they can achieve in 15 seconds,” says Antonio Alegria, head of AI at OutSystems. “And curiously, we’ve seen developers saying … that it’s actually helped them with repetitive strain injuries.”
In the latest episode of The AI Show, we talk to Alegria about smart systems helping developers write software. OutSystems’ low-code platform uses hybrid AI, combining probabilistic deep learning algorithms with configurable hard-coded constraints to predict what developers need next and present that visually.
Trained on “over 15 million instances of code patterns,” the software apparently helps developers reduce cognitive load and focus more on what they want to build, rather than how to specifically actualize it. And if they have to type less, they can reduce repetitive strain on their wrists, elbows, and fingers.
Listen to the podcast here:
“In many cases where the algorithms have learned specific patterns, you almost can just go, ‘Okay, next, next, next, next, next,'” Alegria says. “It’s very powerful.”
To build the platform, OutSystems makes extensive use of graph neural networks and technologies like TensorFlow. But this isn’t an integrated development environment that you can plug multiple programming languages into: It’s a proprietary system for building business applications, which limits its applicability.
And there are additional challenges, because presenting the next best option to developers isn’t always easy. “What we found is that sometimes we have the AI capability to do the right predictions with high accuracy, but it’s not always easy to find the best user experience to expose this,” says Alegria.
Subscribe to The AI Show on your favorite podcasting platform to get the full story: