Modulate raises $25M to detect deepfake voices and read customer intent

Modulate has raised $25 million to expand a voice intelligence platform that runs more than 100 small models for transcription, emotional analysis, deepfake detection and policy enforcement.

Future Ventures led the round, with Hyperplane and Lakestar participating. PitchBook data indicated the Boston-based startup had previously raised $41 million at a $170 million valuation.

The round sits at the intersection of two crowded investor bets in voice AI: making synthetic voices sound more human, and detecting what humans actually mean when they speak. Modulate also competes with companies building defenses against deepfake calls, a growing concern as cloning a voice has become trivially easy.

Mike Pappas and Carter Huffman founded the company in 2017 after meeting as MIT physics undergraduates. Modulate started with voice modulation for gaming, then moved into voice-based moderation tools. The arrival of voice AI models pushed the company toward its current focus: detecting AI-generated audio and analyzing the intent behind a person's words.

"Our insight into the voice AI space is that a lot of folks are doing transcription, but there's not really any capability out there that gets the full nuance and full understanding of a conversation, which is so important when you're talking to another human being," Huffman said on a call with TechCrunch.

Modulate splits its models into two categories. Signal extraction models understand vocal emotion, tone, language and synthetic voice determination. Analysis and detection models examine intent: what a customer is trying to say, whether a caller is violating rules, or whether they are trying to scam the receiver.

The small-model architecture carries practical advantages. Huffman said the company does not need specialized hardware or massive compute, a meaningful factor as token bills climb. The setup also makes it easier to train models with new capabilities, add them to the fleet, and have an orchestrator call them when needed.

Modulate serves a varied customer base, but it specializes in deepfake detection and alerting organizations like call centers to possible scams. It also monitors how AI agents respond to customers, assesses call quality, and checks that AI follows compliance rules in regulated industries. The company often sits beside a company's existing voice stack purely to analyze calls.

As enterprises adopt AI-powered customer service, understanding why a call succeeded or failed has become critical. Huffman argues that basic sentiment analysis misses the mark.

"I think when companies think of emotion analysis, they think if the customer was neutral or positive, the call was a success, and if the customer was negative, the call was a failure. But actually, many times people will be polite even to, like, AI agents or bots. Right. And they won't come across as angry, but they'll be very dissatisfied," he said.

Modulate can give enterprises granular data on customer intent and response beyond that surface-level read. The company said its technology is also being used to monitor cyberattacks conducted through voice calls.

The startup currently employs 40-45 people and plans to add 10 more in the coming months to bolster model building. Modulate is also working on expanding on-premises and on-device deployment options for customers that need stronger privacy guarantees.

The bet for investors is that as voice agents proliferate in regulated industries, the analysis layer around those agents becomes as valuable as the agents themselves.