Abhimanyu Sikarwar

AI Voice Lead Qualification Agent

1 min read · 193 words

Overview

An agentic AI system built with MCP (Model Context Protocol) and LangGraph that calls sales leads on its own, runs a qualification conversation over voice, and scores each lead by interest level. The sales team only talks to people who already sound like buyers.

Tech stack

How it works

The system takes a list of sales leads and places voice calls using Sarvam AI text-to-voice. During each call, the LangGraph agent asks about needs, budget, timeline, and intent, and adapts its follow-up questions to what the lead says.

After each call, the agent scores the lead’s interest and writes a structured qualification summary. The sales team gets a list ranked by likelihood to convert, with the transcript attached to each entry.

Features

Tags: #MCP #LangGraph #Python #AI #Agents #Sarvam AI #WebSocket