Boston-based voice intelligence startup Modulate has raised $25 in new funding for a platform that uses a suite of miniature models to provide enterprise transcription, sentiment analysis, music detection with deepfakes and AI, and policy enforcement for voice agents in regulated industries.
The funding follows a popular trend among investors in the growing voice AI industry: backing companies that are trying to make AI voices sound more human-like.
It also rivals other companies that are trying to analyze human conversations to find out the intentions behind them, as well as companies that are trying to protect people and businesses from deepfake calls because voices can be easily cloned.
Modulate’s new funding was led by Future Ventures, with participation from Hyperplane and Lakestar. Prior to this round, the startup had raised $41 million in funding at a valuation of $170 million, according to PitchBook data.
The startup was founded in 2017 by Mike Pappas and Carter Huffman, who met as undergraduate physics students at MIT. In its early days, the company focused on providing audio modulation for gaming. But then they started focusing on voice-based moderation tools.
With the launch of voice AI models, the company is now focused on detecting different types of AI voice production and analyzing the intent behind human words.
“Our insight into the voice AI space is that a lot of people are doing transcription, but the ability to get the nuances and full understanding of the conversation doesn’t really exist, which is really important when you’re talking to another human being,” Huffman said on a call with TechCrunch.
The company currently runs over 100 models. These models are broadly divided into two sections. A signal extraction model for understanding voice emotion, tone, language, and synthetic speech decisions. An analysis/detection model that looks at intent, such as what the customer is trying to say, whether the caller is violating rules, or trying to deceive the recipient.
Huffman said the company runs a small-scale model that doesn’t require specialized hardware or large amounts of computing, which could become important if token charges rise. Additionally, it is easier for enterprises to train models with new capacity, add them to lots, and call them to the orchestrator as needed.
Modulate has a diverse customer base, but specializes in deepfake detection and alerting organizations such as call centers to potential fraud. It also monitors how AI agents interact with customers to assess call quality and ensures that AI follows compliance rules in regulatory areas. Because of these products, the Module is often placed next to the voice stack used by enterprises just to analyze calls.
As more companies implement AI-powered customer service, it becomes important to know why customer calls are successful or unsuccessful. In that case, it’s important to go beyond basic analysis to assess customer intent and response. Huffman said Modulate can provide companies with detailed data around that.
“When companies think about sentiment analysis, I think they think that if the customer is neutral or positive, the call is a success, and if the customer is negative, the call is a failure. But in reality, people are often polite to AI agents and even bots, and that’s true. And they won’t look angry, but they will be very unhappy,” he said.
The company said its technology is also used to monitor cyberattacks through voice calls.
The startup currently has 40 to 45 employees and aims to add another 10 people in the coming months to ramp up model building. Modulate is currently working on enhancing its on-premises and on-device deployment capabilities to increase privacy.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
