Awaaz AI runs multilingual voice agents for financial-services workflows over phone calls and messaging channels. The agents handle loan origination and eligibility checks, KYC (know your customer) verification, collections, and sales conversations, and can hand off to WhatsApp or SMS mid-call to collect documents.
Conversations run in eight or more languages, including vernacular and mixed-language speech such as Hinglish, and are personalised from the customer's transaction history through CRM and customer-data-platform integrations. Calls are carried on Awaaz's own telephony stack, which the company sizes at 10 million calls a day, and a WebSocket media stream exposes call audio so an agent pipeline such as Pipecat can host the conversation.
An analytics dashboard reports on call outcomes, and billing is per conversation minute. Awaaz states a speech-recognition accuracy above 95% for its target languages.
Features
- Finance-domain agents: loan origination, KYC verification, collections, and sales calls
- Multilingual speech: eight or more languages with mixed-language conversations
- Channel switching: moves from a voice call to WhatsApp or SMS to collect documents
- Telephony stack: in-house calling infrastructure with a WebSocket media-stream interface for agent frameworks
- Personalisation: pulls transaction history through CRM and CDP integrations
- Analytics dashboard: reporting on engagement and call results
- Usage billing: credits charged per conversation minute
