With the ongoing effects of the pandemic, digital-based communications are an absolute necessity. In the spring, NVIDIA (NVDA -6.62%) CEO Jensen Huang announced upcoming software development platforms aimed at AI-based communications and online recommender systems. Dubbed Jarvis and Merlin, NVIDIA's new platforms are now available to software developers using AI to build better services for the next phase of the digital age, Huang announced today. 

Someone in a suit holding a tablet. A brain illustrated with electrical connections hovers above the screen.

Image source: Getty Images.

Jarvis is named after the natural language communications system built by the character Tony Stark in the Marvel Cinematic Universe. Much like the fantasy version of AI software, NVIDIA's iteration is designed to recognize human speech and reply in a natural, conversational manner. 

Every day in the U.S. alone, NVIDIA says, call centers handle an average of 200 million calls, and telemedicine enable 2.4 million virtual-healthcare visits. Jarvis is already at work in cloud-based call centers, easing the burden of daily calls, and has improved the accuracy of voice transcription on medical visits. Digital wallet and payments company Square (SQ -0.30%) was also an early adopter of Jarvis to build a virtual assistant that can schedule appointments.  

Merlin improves the ability of web-based recommendations. Current systems are clunky. Ever get a shopping recommendation based on what you recently viewed online, or a video or music recommendation based on a recent playlist? That software is simply based on recent behavior. NVIDIA's Merlin aims to upgrade these recommender capabilities to something more intuitive and relevant to the user. 

NVIDIA notes that recommendations account for as much as 30% of revenue on some e-commerce sites, so even a small improvement could lead to dramatic business results. Merlin helped Chinese social media giant Tencent Holdings (TCEHY -2.78%) reduce the amount of AI-based algorithm training time from 20 hours down to just three, dramatically increasing the deployment of recommender systems on the Tencent platform.