Our AI model will go beyond words, understand emotions

Bengaluru-based Gnani.ai is working on a model to bring emotionally intelligent, real-time voice AI to the country's vast linguistic landscape.

- 

Selected under the IndiaAI Mission to build a sovereign multilingual voice AI foundational model, the startup is working on a voice-to-voice large language model (LLM).

Co-founder and CEO Ganesh Gopalan explains to Ayanti Bera how the company plans to revolutionise human-AI interaction through its model.

What is your vision for the future of voice AI in India?

Our vision is to break down language barriers and enable real-time, emotionally intelligent communication across the country's diverse linguistic landscape.

We aim to support 22 languages and numerous dialects, making technology accessible for everyone, whether it's for customer service, education, or citizen engagement.

We believe voice is the most natural interface, and if we can understand not just the words but also the emotions behind them, we can truly elevate how AI interacts with people.

Why is emotion recognition important in voice AI?

In India, tone, pitch, and intonation can significantly alter the meaning of a sentence.

Current voice systems that convert speech to text often lose these subtle cues.

We are building a system that preserves and interprets emotional signals to provide more accurate, context-sensitive responses.