Registration is open for 2026 MARKETPLACE Wisconsin.

Tuning out the noise: How Antimattr is advancing voice-first AI

November 20, 2026
Share This Story:

Antimattr foundersColumn by Sridipto Ghosh, Co-Founder and CEO of Antimattr

Voice is becoming one of the most important ways people interact with artificial intelligence (AI), but background noise can make it difficult for AI tools to understand what we’re saying. At Antimattr, we’re developing voice-first technology that helps people communicate with AI more clearly—whether they’re in a busy office, on the go, or working in a noisy environment.

I’m originally from Bangalore, India, and studied engineering at the Indian Institute of Technology. Before starting Antimattr, I co-founded an enterprise-focused company with a childhood friend, which grew to nearly $250,000 in revenue within its first 18 months. That experience taught us a lot about building a business, but we knew we wanted to take on something even more ambitious. We saw an opportunity in AI and how people could interact with it, and that became the starting point for Antimattr.

Today, we are building voice-first AI hardware designed to help people do more with their voice. Our products include a smart ring and a pair of earphones, supported by an app that allows users to control their devices and access information. The goal is to make it easier to capture thoughts, manage work, and interact with AI agents without having to stop what you’re doing and type everything out.

One of the biggest challenges we identified is that while voice models have become very capable, they don’t always work well outside of quiet environments. Background noise can interfere with the input, which means even the best AI technology may not deliver the results people expect. We realized that instead of trying to improve the AI model itself, we needed to solve the problem at the interface level: helping the technology hear the user clearly in the first place.

That’s where our hardware comes in. Our earphones use specialized microphones, sensors, and software to isolate the user’s voice, filtering out surrounding noise during calls and dictation. The ring provides another way to interact with technology, allowing users to tap a button, capture a thought, or control an AI agent through voice. We are designing these products for people whose work is heavily centered around communication—including engineers, founders, entrepreneurs, agency owners, doctors, and lawyers.

The early response has been encouraging. We’ve already received nearly 1,200 preorders, and we’re also seeing opportunities beyond individual consumers. We believe our voice-isolation technology could be especially valuable for businesses developing AI agents that need to operate in noisy environments.

Our experience in Wisconsin through the global accelerator program by gener8tor and supported by the WEDC helped us better understand the U.S. market and build relationships that could support our growth. The gener8tor team was incredibly helpful with everything from navigating visas and relocation to connecting us with mentors and potential partners.

One of the most valuable opportunities that came out of the program was recognizing the potential for enterprise partnerships in Wisconsin. The state has a strong manufacturing presence, and through the network we developed, we were able to connect with factories that could potentially become customers or partners. It helped us see how our technology could solve real challenges in industries where noise is part of the everyday work environment.

Wisconsin could also play an important role in our long-term manufacturing plans. We are preparing to produce our first batch of devices in Shenzhen, China, which will give us an opportunity to learn through the initial manufacturing process and refine the product. As we grow, we hope to develop our own manufacturing capabilities and have greater control over production. Given Wisconsin’s manufacturing expertise, it is a place we see as a strong potential fit.

Looking ahead, our immediate focus is on building the best voice-isolation technology possible and making it a seamless part of how people work. But we see that as just the beginning. The next step is to make AI agents more accessible to everyone. Today, using or building an AI agent can still require some technical knowledge. We want to make that experience as intuitive as using a mobile app, so anyone can create and run their own agents.

One piece of advice from a gener8tor mentor has really stayed with me: “A lot of venture is not teachable, but is learnable.” I think that’s true of entrepreneurship in general. You can read about building a business or learn from other people’s experiences, but ultimately, you have to go out and build something yourself. You learn by doing, making mistakes, and figuring things out along the way. That’s the mindset we’re bringing to Antimattr as we continue to grow.

Related Posts

Go to Top