Founder Stories

What if every Indian were understood in their own language?

‍

An Indian voice AI startup trained on 50,000 hours of real Indian speech has raised ₹40 Cr led by Ronnie Screwvala, with Bajaj Finserv, Ujjivan and Jana Small Finance Bank already on board.

Antler India

Company

October 14, 2026

Share article

India's contact centres take billions of calls a year in 22 languages and more than 40 dialects. In banking, insurance and healthcare, a misheard word is more than an annoyance. It can mean a wrong loan, a missed claim, a compliance breach or just sheer disrespect. Most voice AI was trained on American English, and stumbles the moment a caller switches from Marathi to Hinglish mid-sentence.

Raoul, a second-time founder who earlier built a mobile gaming studio, along with his brother Jai, a software engineer from the University of Michigan, decided to build the whole thing themselves. Speech recognition, voice synthesis and language models, trained on more than 50,000 hours of real Indian speech and deployed inside each customer's own private cloud. When it comes to speech in Indian languages, Navana is 30% more accurate than anyone else.

Bajaj Finserv, Ujjivan and Jana Small Finance Bank now trust Navana with their customers' voices. A farmer in Vidarbha or a shopkeeper in Madurai speaks to their bank as naturally as anyone in Mumbai, and is served just as well. Recently, Navana raised 40 Cr led by Ronnie Screwvala to accelerate growth and become the de-facto voice AI for Bharat.


FOUNDERS

1. Raoul Nanavati, Co-founder & CEO

2. Jai Nanavati, Co-founder & CTO

‍

ANTLER RESIDENCY —LAUNCH YOUR STARTUP

Antler backs exceptional founders to go further, faster.

Apply now

MORE INSIGHTS

Founder Stories
November 2, 2026

From First Customers to Global Markets

Applications for Antler Australia's residency open Monday, 2 November, 2026. This is what the eight weeks look like, told through the founders who have been through them, from a first-time founder leaving a bank to a company building drugs for advanced brain disease.

Australia
Founder Stories
October 14, 2026

Why we invested in Cloop

The modern knowledge worker doesn't have a productivity problem. They have an attention problem.

UK
Founder Stories
October 14, 2026

Why we invested in Marco

Every time a robotics team swaps a sensor, shifts a mount or moves into a new building, someone has to re-tune its navigation by hand. That often takes one to four days per robot. Larger teams have whole departments for this, and one autonomous driving company we spoke to runs a 24-person sensor calibration team. The founders estimate that only around 200 people worldwide have deep experience combining inertial sensors with machine learning, and most of them work at Apple, Google, Meta, Qualcomm and Huawei.

UK

BUILD YOUR STARTUP
—WITH ANTLER

Turn your vision into a world-changing company. Apply to Antler and start building alongside a global network of founders and investors.

Apply now
Woman standing in front of Antler cohort