Customer Support Voice Assistant AI

I am building a conversational AI that can serve customers through voice, not just text. The assistant must act as a full-service front line, able to:

• Answer our most common FAQs with friendly, accurate responses
• Handle complaints empathetically and log them with proper categorisation and escalation tags
• Guide callers through order placement, status checks, and basic modifications, then push the data into our existing order-management API

Fluency in both English and key Indian languages is essential so every caller receives the same high-quality experience. Natural, human-sounding speech synthesis and rapid intent recognition are priorities; dead air or obvious “bot pauses” are deal-breakers.

I will supply current FAQ content, complaint workflows, and API documentation for orders. You will turn these into a production-ready voice skill: training the NLU, configuring TTS/STT engines, setting up call flows, and deploying to the cloud platform we agree on (AWS Connect, Dialogflow CX, or a comparable stack—happy to take your recommendation).

Acceptance is complete when:

1. A test phone number routes to the assistant.
2. All three task types above pass scripted and unscripted test calls with ≥95 % accuracy.
3. Logs show correct language selection and error handling.
4. Handover to a live agent works smoothly when the user says “representative.”

Once live, I’d also like simple dashboards for usage and accuracy so we can iteratively improve the model. Let’s create a voice experience our customers actually enjoy.

example sites: https://telleo.com.au/

telleo.ai

https://voizematic.ai/

demo is required