BOSTON, MA / ACCESS Newswire / November 10, 2026 / Modulate, the frontier audio AI company, today announced new investments from Twilio Ventures and technology entrepreneur Paul English, co-founder of Kayak, alongside significant advancements to its Velma audio intelligence platform. The updates enable enterprises to identify AI callers in real time and gain greater visibility into the performance of AI voice agents. The investments follow Modulate's recently announced $25 million in funding led by Future Ventures, and come as the company expands its technology to address growing demand for enterprise.
Already putting that funding to work, Modulate has released new updates to its Velma audio intelligence platform designed to address two converging trends: a growing flood of calls placed by personal AI voice agents, such as Meta's Muse, and the continued rise of synthetic voice and deepfake threats. The updates also give enterprises greater visibility into the increasingly complex conversations taking place between humans and AI. Modulate's latest models have reduced the median time to identify a synthetic voice from 2.25 seconds to 1.6 seconds, while maintaining greater than 99.1% detection accuracy on real-world AI and human voice calls. The technology is designed to identify synthetic callers quickly enough for enterprises to route, flag or respond to AI-generated calls while the interaction is still underway.
"We're reaching a point where AI agents will be calling businesses, talking to customers, and interacting with other AI agents every day," said Carter Huffman, CEO and co-founder of Modulate. "Call operators - human or AI agent - must be able to recognize what's happening in those conversations and respond appropriately while the interaction is still ongoing. We're focused on delivering real solutions to these challenges, ready to deploy at scale and within economics that make sense for businesses."
Faster Synthetic Voice Detection for Real-Time AI Caller Identification
As autonomous voice agents become easier to deploy, enterprises increasingly need to distinguish between human callers and synthetic agents in real time. Contact centers and other voice systems may want to identify AI-generated calls in order to route them differently, apply different policies, track their behavior, or determine whether additional verification is warranted.
Modulate's synthetic voice detection technology analyzes the underlying audio without relying on watermarks from individual model providers, enabling it to identify synthetic speech from a wide range of generation systems, including newer speech-to-speech and full-duplex voice models.
The same capability can be combined with Velma's broader conversation intelligence for fraud and security applications. By analyzing additional signals including intent, manipulation, tone, and conversational behavior, Velma can help enterprises distinguish legitimate AI agents from potentially fraudulent or suspicious interactions.
"As personal voice agents become more common for everyday tasks like booking a dinner reservation, knowing that a voice is synthetic on its own is no longer a useful metric," Huffman said. "An AI agent making a legitimate appointment and a deepfake impersonating an employee may both use synthetic speech. Businesses need enough context to know how to handle each interaction while it's happening, whether they're managing a single phone line or a global customer support operation."
New Audio-Native Signals Expand AI Agent Quality Monitoring
Modulate is also expanding Velma's role as an audio-native observability layer for companies deploying AI voice agents, providing deeper insight into agent performance, conversation quality, and where interactions may be breaking down.
This month, the company added more than a dozen new audio-native signals to Velma, giving developers and enterprises greater visibility into the quality of AI-led conversations. These capabilities can help identify issues such as response timing, audio quality and conversational misunderstandings that cannot be reliably measured from transcripts alone.
Traditional agent analytics largely evaluate what was said. Velma also analyzes the underlying audio and how a conversation unfolds, including emotion, cadence, hesitation, emphasis and changes in conversational behavior. These insights help enterprises evaluate agent performance and identify opportunities to improve the customer experience.
Additional Investment Supports Modulate's Voice AI Momentum
Twilio Ventures and Paul English are the latest investors to support Modulate's next stage of growth. Twilio Ventures' investment brings support from a major communications infrastructure provider as Modulate expands its technology for identifying synthetic callers and monitoring AI voice agent performance. The additional capital will support continued AI and machine learning research, product development, developer adoption, partnerships and expanded deployment options for Modulate's audio-native models.
Twilio Ventures and Paul English join the group of investors supporting Modulate's next stage of growth following the company's recently announced financing led by Future Ventures, with participation from Hyperplane and Lakestar.
About Modulate
Modulate is a frontier audio AI company building technology that understands what is happening in voice beyond the words being spoken. Its Velma platform uses an Ensemble Listening Model architecture comprising more than 100 specialized audio models to analyze signals including emotion, tone, intent, cadence, synthetic speech and conversational behavior.
Modulate's technology has analyzed more than 600 million hours of audio and is used across gaming, fraud and security, healthcare, customer experience, AI agent supervision, trust and safety, and other voice-driven applications. The company's transcription and deepfake detection technologies have ranked #1 on public Hugging Face benchmarks.
For more information, visit www.modulate.ai.
Media Contact
Kristin Canders
Grithaus Agency
kristin@grithaus.agency
SOURCE: Modulate
View the original press release on ACCESS Newswire