A new AI meeting assistant, Krisp, aims to make speakers more easily understood by changing their accent to make it sound North American. The company claims this will make people more easily understood in global meetings. The app works with conferencing platforms including Zoom, Teams, Google Meet and Webex and modifies the speaker’s voice in real time, with just 200ms of processing latency (an imperceptible lag).
Krisp works by modifying the phonemes of speech to make them sound like American English. With AI productivity tools now having more of an impact in meetings, with many people using AI transcription services, this could be seen as the next logical step. But do we really want everyone to sound exactly the same?
Arto Minasyan, co-founder/president of Krisp, knows how challenging it can be to communicate effectively in global business settings with a non-native accent.
“As someone with an accent, I’ve often noticed that people struggle to understand me, even when my English is fluent,” he explains. “This isn’t about bias – it’s simply a reality of communication. In fast-paced meetings, even small misunderstandings can slow down decisions, cause repetition, and reduce overall efficiency. And while improving an accent is possible, it takes years of practice and is incredibly difficult for busy professionals. AI Accent Conversion bridges this gap instantly, ensuring that ideas are heard clearly and conversations flow smoothly, no matter where the speaker is from.”
Krisp currently works for Indian accents, with other accents such as Filipino and South American coming soon. It also assists in meetings with other functions such as generating meeting summaries and providing transcriptions.
A free plan with up to 60 minutes per day is available, and a business plan costs $15 per month.
I see that there may be use cases for Krisp, but it also makes me a little sad to think of a homogenised world where individual accents are no longer. If we are going to embrace this technology, it would also be good if there were more options for different accents to convert from/to.
What do you think of the tech? Let us know in the comments.
