AI Phone Agent Development
Budget: ₹750 – ₹1,250 INR
I need a self-contained voice-based AI that can place and receive calls, answer common customer-service questions, and run simple outbound sales or follow-up campaigns without routing through an existing CRM or VoIP platform. The whole stack—telephony, speech-to-text, language model, text-to-speech, and call logic—must run as a standalone service I can deploy on my own cloud account.
Core scope
• Two conversation tracks: inbound support (order status, basic troubleshooting, FAQs) and outbound promotional or warm-lead calls.
• Natural language understanding good enough to handle unscripted replies, stay on topic, and gracefully hand off to a live agent when confidence drops.
• Simple dashboard or logs so I can review transcripts, call outcomes, and adjust prompts or flows without touching code.
• Secure API endpoint for future integrations, but nothing external is wired in for now.
I’m comfortable if you build on Twilio Voice, Asterisk, or an open-source SIP stack; for the conversational layer, Dialogflow CX, Rasa, or a GPT-powered custom service are all acceptable as long as latency stays low and costs are predictable.
Acceptance criteria
1. I can trigger an outbound call from a CLI or webhook, watch the agent converse, and see the transcript stored.
2. Incoming calls reach the agent, which can resolve at least 80 % of a provided test set of customer queries without escalation.
3. Sales script branch can collect contact confirmation and log a “qualified” or “not interested” flag.
4. Deployment instructions work on a clean Ubuntu LTS VPS.
Show me previous voice-AI work or a short demo snippet in your bid so I know you’re comfortable with end-to-end telephony.
Core scope
• Two conversation tracks: inbound support (order status, basic troubleshooting, FAQs) and outbound promotional or warm-lead calls.
• Natural language understanding good enough to handle unscripted replies, stay on topic, and gracefully hand off to a live agent when confidence drops.
• Simple dashboard or logs so I can review transcripts, call outcomes, and adjust prompts or flows without touching code.
• Secure API endpoint for future integrations, but nothing external is wired in for now.
I’m comfortable if you build on Twilio Voice, Asterisk, or an open-source SIP stack; for the conversational layer, Dialogflow CX, Rasa, or a GPT-powered custom service are all acceptable as long as latency stays low and costs are predictable.
Acceptance criteria
1. I can trigger an outbound call from a CLI or webhook, watch the agent converse, and see the transcript stored.
2. Incoming calls reach the agent, which can resolve at least 80 % of a provided test set of customer queries without escalation.
3. Sales script branch can collect contact confirmation and log a “qualified” or “not interested” flag.
4. Deployment instructions work on a clean Ubuntu LTS VPS.
Show me previous voice-AI work or a short demo snippet in your bid so I know you’re comfortable with end-to-end telephony.
Related categories:
Linux
Asterisk PBX
Voice Talent
VoIP
API Development
SIP
Natural Language Processing
AI Chatbot Development