Navana.ai has raised ₹40 crore in Series A funding to expand voice-agent deployments in banking and financial services. Its proprietary speech stack reportedly supports 12 Indian languages and 45 dialects, with funding also allocated to an AI contact centre and automated conversation-evaluation platform. Inc42 — 7 September 2026
- What changed: Navana is moving beyond transcription toward full conversational workflows, analytics and automated quality assessment.
- Why it matters: Indian voice systems must handle dialects, noisy calls, code-switching, regulatory scripts and low-confidence recognition—not merely translate an English chatbot.
- Practical takeaway: Persist the audio reference, transcript, detected language, confidence, model version, tool calls and human corrections for every voice-agent interaction.
- Analysis: The company’s claimed 100-million-minute processing history and language coverage require independent workload evaluation.
