top of page

Minerva CQ Announces Industry’s First And Most Accurate Italian Speech Recognition on NVIDIA Riva

SUNNYVALE, CA, October 6, 2022 – Minerva CQ today announced Italian language availability for its leading real-time collaborative intelligence platform for the enterprise. The underlying technology marks the first automatic speech recognition (ASR) model for Italian using the NVIDIA Riva software development kit. The model, built in collaboration with Data Monsters, also boasts the industry’s highest accuracy with a word error rate (WER) of under 7%, demonstrably outperforming SOTA results on Mozilla’s Common Voice Italian Benchmark – the global standard for ASR performance.


Companies using the Minerva CQ platform to support Italian-speaking customers can expect to vastly improve their customer experience and agent performance in a much shorter time frame, as tuning and optimization cycles are significantly reduced through the use of best-in-class AI technology. Minerva CQ expects to go live using the Italian ASR model and NVIDIA Riva with a global energy company and major telecommunications provider in early 2023.


“We are excited to bring the power of collaborative intelligence to a broader global market, augmenting our current capabilities with English and Spanish,” said Dr. Cosimo Spera, co-founder and CEO of Minerva CQ. “The collaboration between ourselves, Data Monsters, and NVIDIA demonstrates the power of top-notch data scientists supporting best-in-the-world technology, running on state-of-the-art platforms, releasing products that truly augment the human experience with the power of artificial intelligence.”


“Every customer is delighted to be served in their own language, and with NVIDIA Riva, Minerva CQ customers can now offer responsive services to millions of Italian speakers,” said Hemant Dhulla, general manager of conversational AI at NVIDIA. “The integration of NVIDIA Riva with the Minerva CQ platform accelerates call centers powered by speech AI, helping developers build customized applications in any language.”


“At Data Monsters, we’ve been working with natural language processing for ten years. It’s hard to surprise us (we’ve seen a lot in the area of speech) but frankly we are amazed at how GPU-based speech technologies have advanced in just three years,” said Dan Lesovodski, VP of AI at Data Monsters. “NVIDIA Riva is an excellent platform that drives state-of-the-art results in real, not academic, business-ready products. We’re happy to work with Minerva CQ and NVIDIA, and we believe that speech technologies will gain much wider adoption throughout the consumer and enterprise markets very soon.”


About Minerva CQ

Minerva CQ, the leader in real-time collaborative intelligence for the enterprise, is an AI platform that enables today’s workforce to effortlessly blend with AI during voice and digital conversations to improve agent and customer experiences. In the contact center, the Minerva CQ platform provides real-time agent guidance, automated QA and training suggestions, and business intelligence. In the broader enterprise, Minerva CQ augments agents supporting field technicians with knowledge surfacing so they can tackle complex issues, repairs, and logistics.


Press contact:


bottom of page