Banking Voicebot for transcation with Voice Authentication
Budget: ₹750 – ₹1,250 INR
I need a conversational voice assistant that lets our retail-banking customers hear their recent transaction history after passing voice-print verification. The flow is simple: the caller greets the bot, the system confirms identity through speaker recognition only, then pulls and reads the latest transactions from our back-end.
Core scope
• Functionality: Transaction-history inquiry (no balance checks or bill pay in this phase).
• Authentication: Voice recognition — no PINs, no security questions.
• Languages: English, Telugu, Tamil, with the entire dialogue staying in whichever language the user starts with.
What I’d like from you
• Design the conversational flow and intents.
• Implement biometric speaker verification; propose any SDK or cloud service that meets banking-grade security.
• Integrate a sample REST/JSON transaction API (Swagger spec supplied).
• Provide fully commented source, deployment notes, and a lightweight test script so my QA team can run end-to-end checks.
Acceptance criteria
1. Voice auth completes in <3 s and rejects impostors.
2. ≥95 % intent accuracy across all three languages in quiet-line tests.
3. Correct transactions read back—no data leakage or mix-ups.
4. Code builds and runs in my AWS sandbox without extra, undisclosed licence fees.
If you have prior experience with voicebots in regulated industries, please highlight that along with your proposed tech stack and timeline. I’m ready to move once we lock the architecture and milestones.
Core scope
• Functionality: Transaction-history inquiry (no balance checks or bill pay in this phase).
• Authentication: Voice recognition — no PINs, no security questions.
• Languages: English, Telugu, Tamil, with the entire dialogue staying in whichever language the user starts with.
What I’d like from you
• Design the conversational flow and intents.
• Implement biometric speaker verification; propose any SDK or cloud service that meets banking-grade security.
• Integrate a sample REST/JSON transaction API (Swagger spec supplied).
• Provide fully commented source, deployment notes, and a lightweight test script so my QA team can run end-to-end checks.
Acceptance criteria
1. Voice auth completes in <3 s and rejects impostors.
2. ≥95 % intent accuracy across all three languages in quiet-line tests.
3. Correct transactions read back—no data leakage or mix-ups.
4. Code builds and runs in my AWS sandbox without extra, undisclosed licence fees.
If you have prior experience with voicebots in regulated industries, please highlight that along with your proposed tech stack and timeline. I’m ready to move once we lock the architecture and milestones.