Modulate secures $25M for enterprise voice intelligence platform
Boston-based voice intelligence startup Modulate has raised $25 million in new funding for an enterprise platform that combines transcription with emotional analysis, deepfake detection, AI music detection and policy enforcement for voice agents in regulated industries. Future Ventures led the round, with participation from Hyperplane and Lakestar.
The company operates more than 100 models across its product suite. Before this round, PitchBook data indicated that Modulate had raised $41 million at a $170 million valuation. Founded in 2017 by MIT physics undergraduates Mike Pappas and Carter Huffman, the startup initially developed voice-modulation technology for gaming before shifting towards voice moderation and, later, analysis of AI-generated audio and conversational intent.
Small models orchestrated for specific voice tasks
Modulate groups its models into two broad categories. Signal-extraction models assess vocal emotion, tone and language and determine whether a voice is synthetic. Analysis and detection models evaluate intent, including what a customer is trying to communicate, whether a caller is breaking rules, or whether a caller may be attempting a scam.
Huffman said the company’s approach is intended to capture nuances in a conversation that transcription alone does not provide. The company uses smaller models rather than relying on a single large system, with an orchestrator selecting models when they are needed.
That architecture, Modulate says, reduces the need for specialised hardware and large volumes of compute. It also allows the company to train and add models with new capabilities. This is relevant as enterprises weigh the operating cost of voice AI systems and seek analysis that goes beyond converting speech into text.
Call-centre monitoring, scams and AI-agent compliance
Modulate is used for deepfake detection and for alerting organisations, including call centres, to possible voice scams. It also monitors how AI agents respond to customers, assesses call quality and checks whether those agents follow compliance rules in regulated settings.
The platform generally sits alongside a company’s existing voice stack to analyse calls. Its purpose is not limited to classifying customer sentiment as positive, neutral or negative. Huffman noted that callers may remain polite when speaking to an AI agent while still being dissatisfied, making intent and response analysis important to understanding whether a call succeeded.
The company also said its technology is used to monitor cyberattacks conducted through voice calls. Modulate has 40 to 45 employees and plans to add 10 staff in the coming months to strengthen model development.
Privacy becomes part of the deployment roadmap
Modulate is working to expand on-premises and on-device deployment options, positioning privacy as a focus for customers processing voice data. For businesses deploying AI-powered customer service, the practical implication is to evaluate voice systems not only for transcription accuracy, but also for synthetic-audio detection, intent analysis, agent oversight and deployment controls that match compliance requirements.