Modulate raises $25M for voice deepfake and scam detection tools

In short: Modulate, a Boston startup, raised $25 million to expand its software that analyzes voice calls and helps detect deepfake audio, fraud, and scams.

What happened

Modulate said it raised $25 million in new funding for its voice analysis platform. The round was led by Future Ventures, with participation from Hyperplane and Lakestar.

The company sells tools that can transcribe calls, which means turning speech into text, and analyze things like tone and emotion. It also aims to spot synthetic audio, which is audio made by AI, including deepfake voices (a fake voice meant to sound like a real person).

Modulate says it runs more than 100 smaller AI models (think of them like many specialized mini tools in a toolbox, rather than one giant tool). These models look for signals like language, tone, and whether a voice sounds computer-generated. Other models try to figure out intent, like whether a caller is trying to scam someone or break a company’s rules.

Modulate was founded in 2017 by Mike Pappas and Carter Huffman. It first focused on changing voices for gaming, then shifted toward voice moderation, and now focuses on detecting AI-generated audio and analyzing conversations. The company has about 40 to 45 employees and plans to hire about 10 more, and it is working on more on-premises and on-device options (meaning the software can run inside a company’s own systems or directly on a device, which can improve privacy).

Why it matters

Phone scams and fake voice calls are becoming easier to pull off because voices can be cloned. Tools like Modulate’s are trying to help call centers and other businesses tell the difference between a real customer and a fake, and to understand why a call went well or badly, not just whether someone sounded happy.

Source: TechCrunch AI