AI-powered audio company Krisp has unveiled a new feature designed to modify accents in real time, allowing Indian English speakers to sound more like native American English speakers during calls.
The company announced on Wednesday that this accent conversion tool is now available in beta on its desktop app, with plans for broader deployment in the future. According to Krisp, the technology retains the speaker’s natural voice while altering specific phonemes to better match American pronunciation. Users can enable this feature either before or during calls.
Addressing Communication Barriers
Krisp’s co-founder, Arto Minasyan, explained that the inspiration for this innovation came from his own experiences. “Even though I speak English fluently, people sometimes struggle to understand my accent. We realized that modifying accents could enhance communication. That’s what led us to develop this feature over the past two years, and now we’re ready to launch it in beta,” he said.
The feature is still a work in progress. Initial testing revealed that the modified voice didn’t always sound entirely natural and occasionally omitted words. Krisp acknowledged these shortcomings, attributing them to the beta phase and promising improvements as the AI model continues to evolve.
Impact on Businesses and Call Centers
The company claims that early testing of the accent modification tool in enterprise environments has already shown promising results. Businesses using the feature reportedly experienced a 26.1% increase in sales conversions and a 14.8% rise in revenue per booking.
Krisp chose to focus on Indian accents first due to the country’s large representation in the global workforce, particularly in STEM and customer service roles. Future updates will expand support to additional accents, including Filipino.
How the AI Model Works
To build this technology, Krisp trained its AI on thousands of speech samples representing diverse accents and dialects. The company also incorporated data from its meeting assistant, ensuring user consent in the process.
One of the key benefits of the system is that it doesn’t require prior training on an individual’s voice. Instead, it analyzes speech in real time and applies modifications instantly, eliminating the need for pre-recorded samples.
Expanding to Mobile and Browser Integration
Krisp, which last secured funding in 2021, has ambitious plans for expansion. This year, the company aims to release iOS and Android apps to bring its technology to in-person meetings. Additionally, a Chrome extension is in development to enhance Google Meet compatibility.
While Krisp isn’t the only player in this space—other startups, such as Sanas, have deployed similar AI-driven solutions in call centers—the company believes its real-time approach will set it apart.
