Boston-based voice intelligence startup Modulate has raised $25 million in funding led by Future Ventures, with participation from Hyperplane and Lakestar, to develop its platform. According to PitchBook data, the company had raised $41 million at a valuation of $170 million before this round.
Founded by Mike Pappas and Carter Huffman, who met in 2017 while studying physics at MIT, Modulate initially developed voice-changing technology for games. The company later shifted its focus to voice-based moderation; today, it focuses on detecting AI-generated voices and analyzing intent in conversations.
Modulate provides businesses in regulated industries with transcription, sentiment analysis, deepfake and AI-generated music detection, and services for monitoring voice agents’ compliance with policy rules. The company operates more than 100 small models. These models are divided into two groups:
- Signal extraction models that identify vocal emotion, tone, language, and synthetic voices.
- Analysis and detection models that examine what the customer intends to say, whether the caller is violating rules, or whether they are attempting fraud.
According to Huffman, using small models reduces the need for specialized hardware and high computing power; it also makes it easier to add new capabilities to the models. Modulate is particularly working on deepfake detection and alerting call centers to potential fraud. The system also monitors the quality of responses AI agents give to customers and their compliance with regulations. The company also says it is used to track cyberattacks through voice calls.
The startup has 40-45 employees and plans to hire 10 more people in the coming months to strengthen its model development team. Modulate is expanding its in-house and on-device execution capabilities to enhance privacy.
Why it matters
The significance of this investment lies in treating voice communication as an area that needs to be monitored not only for user experience but also for security, fraud, and regulatory compliance. Modulate’s focus on regulated businesses has direct implications for call centers, organizations using voice agents, and teams overseeing these services. The structure consisting of more than one hundred small models makes it possible to examine different voice signals and speech intent separately, while the effort to enable in-house and on-device operation brings the question of where data is processed to the forefront. The new resources will support model development aimed at detecting deepfakes, synthetic voice, and cyberattacks carried out over the phone. The open question is how accurately these systems perform in real calls and how they will be integrated into organizations’ compliance processes.
Term: agent
An AI agent is software that calls tools and carries out multi-step tasks to achieve a goal, rather than simply generating a single response.