Is India ready for an AI wearable that is always listening?

When cousins Dhananjay Yadav and Aryan Yadav were hatching their AI wearables startup, NeoSapien, a Bengaluru café doubled as their laboratory.

Before investor decks and funding announcements, they tested their idea – that the next interface for AI would not involve typing on a screen – in full public view.

The pendant-like device, placed casually on a table at Araku Coffee on busy afternoons, eventually started getting attention. Founders and investors drifting through the café would pause, ask questions, try it on, and often post about it on X. Its prototypes had been assembled by Dhananjay and Aryan themselves, soldered together at home before being brought out for real-world feedback.

That device evolved into Neo 1, an AI wearable that listens to conversations, transcribes them, and constructs what Dhananjay describes as a “deep context layer” around its user.

Therefore, the “prompts” are more than what gets typed into a chatbot. The data Neo 1 uses includes the full arc of a person’s daily interactions, including conversations with family, friends, colleagues, and even doctors. He says it’s only when that is included that AI can move from responding to queries to proactively helping with decisions.

“India is a voice-first market,” Dhananjay points out. “If you build for voice properly, especially in multiple languages, you unlock a very different user base.”

He explains that voice-first AI lowers barriers in two key ways: It removes the need for fast typing or strong English skills, and it works even when people are on the move, like during field visits or informal meetings.

Quick daily tasks will move to AI wearables. It’s just evolution.

The premise, Dhananjay argues, is straightforward. AI understands maybe 5% of a user’s conversations without additional context, he tells Tech in Asia.

“If it is going to be truly powerful, it needs 100% of your context,” Dhananjay adds.

Neo 1 recognizes when someone is talking using voice activity detection and identifies who that is via speaker diarization, an AI-powered audio process.

In ambient environments, however, the device can’t completely avoid capturing nearby conversations. But the company clarifies that because of privacy concerns, it deletes the audio files as soon as they are processed. It also says it doesn’t link third-party voices to identifiable data.

NeoSapien is now building an application layer on top of that captured context. The idea is that everyday micro-tasks – such as ordering food, booking a cab, sending emails, taking notes, and managing schedules – could shift from smartphones to AI wearables, Dhananjay shares.

Always-on AI

Hardware dreams, geopolitical realities

Hype, velocity, and regulation