Boston-based voice intelligence startup Modulate has raised $25 million in new funding to expand its platform for transcribing, analyzing, and moderating voice conversations. The round, led by Future Ventures with participation from Hyperplane and Lakestar, arrives as enterprises rush to deploy AI voice agents and simultaneously defend against AI-generated deepfake scams.
The company runs more than 100 small models that perform two core functions. Signal extraction models parse vocal emotion, tone, and language while flagging synthetic voices. Analysis and detection models then examine intent - what a customer is really trying to say, whether a caller is violating rules, or whether a scam is in progress.
Prior to this round, Modulate had raised $41 million at a $170 million valuation, according to PitchBook data. The startup was founded in 2017 by Mike Pappas and Carter Huffman, who met as undergraduates in MIT's physics program.
From gaming moderation to voice intelligence
Modulate began with voice modulation tools for online gaming before shifting toward moderation. The explosion of voice AI models pushed the company further, toward detecting different types of AI-generated audio and analyzing the intent behind spoken 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.
The company's reliance on smaller models means it avoids the specialized hardware and heavy compute costs that larger models demand. That architecture also lets Modulate train new models for specific capabilities and deploy them through an orchestrator that calls each model as needed.
Deepfake detection and compliance enforcement
Modulate's customer base spans call centers, regulated industries, and organizations needing deepfake detection. The platform alerts clients to possible voice scams and monitors how AI agents handle customer interactions, assessing call quality and checking that AI follows compliance rules. The technology also detects cyberattacks conducted through voice calls.
For enterprises adopting AI-powered customer service, understanding why a call succeeded or failed has become critical. Huffman argued that surface-level sentiment analysis misses the point.
"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. And they won't come across as angry, but they'll be very dissatisfied," he said.
Privacy-first deployment and team growth
The startup employs 40 to 45 people and plans to add 10 more in the coming months, primarily to strengthen model development. Modulate is also working to expand its on-premises and on-device deployment options, a move aimed at organizations with strict data privacy requirements.
For communications and marketing professionals managing customer experience strategy, the shift toward granular voice analysis changes how teams measure satisfaction. Modulate's approach suggests that tone alone is a weak signal - intent and unspoken dissatisfaction matter more. Professionals overseeing AI Customer Service Training or working in regulated sectors may find this level of analysis increasingly relevant as voice AI adoption accelerates. Those navigating compliance requirements can explore AI Regulatory Compliance Courses designed for specialists in heavily supervised industries.
Why this matters for creatives, marketing, PR, and communications professionals
Voice AI is reshaping customer touchpoints faster than most brand guidelines can adapt. When AI agents handle calls, the quality of those interactions reflects directly on the brand. Tools that analyze not just what was said but what was meant - and whether the voice on the other end was even real - will influence how teams brief voice agents, write escalation scripts, and measure campaign outcomes. The takeaway: sentiment scores won't cut it anymore. Granular intent data and deepfake detection are becoming table stakes for protecting brand reputation in voice channels.
Your membership also unlocks:
