Key facts
- Gnani AI is a Bengaluru-based conversational AI startup
- New model focuses on speech-to-text for Indian languages and accents
- Target sectors include banking, financial services, and insurance
- Addresses underperformance of global models on Indian speech patterns
Deccan Herald reports that Gnani AI, a Bengaluru-based artificial intelligence startup focused on conversational and voice AI, has launched its latest speech-to-text model. The company has built its business around serving enterprise clients — particularly in banking, financial services, and insurance — with AI-driven voice solutions including automated call centre tools and transcription services.
Gnani AI's focus on Indian languages and the distinctive accents and code-switching patterns of Indian speakers addresses a well-documented gap in global speech AI, where models trained predominantly on Western English data tend to perform poorly on Indian English and vernacular languages. A model purpose-built for this context could deliver meaningfully better accuracy for businesses handling high volumes of customer calls in India.
The launch reflects the growing maturity of India's homegrown AI ecosystem, with startups increasingly moving beyond reselling global models to building specialised systems for local needs. For Indian businesses that rely on voice-based customer interaction — from banks to insurers to telecom firms — improvements in speech-to-text accuracy directly translate to better customer service and lower operational costs.