Claude AI Developer for Voice & CRM

Summary

Project Overview

I am looking for an experienced AI/API developer to build an automated workflow using Anthropic’s Claude. The system must process both incoming voice calls (via speech-to-text) and live chat messages, extract key data, and automatically log structured summaries into our CRM.
Key Requirements

1. Voice AI Integration

Transcribe inbound/outbound sales or support calls using a provider like Twilio, Deepgram, or AssemblyAI.

Pass the transcribed text to Claude (Anthropic API) to identify:

Action items

Customer sentiment

Key dates/commitments

Return a structured JSON summary.

2. Chat AI Integration

Connect Claude to a live chat widget (e.g., Intercom, Drift, or custom webhook).

Enable Claude to respond to common customer queries or summarize agent-customer chats.

Ensure the same data extraction (action items, intent, urgency) as voice.

3. CRM Automation

Automatically push Claude’s structured output from both voice and chat into our CRM ([e.g., Salesforce, HubSpot, Pipedrive, Close, or custom API]).

Fields to update per interaction:

Call/chat transcript summary

Follow-up tasks & due dates

Lead stage or ticket priority

Avoid duplicates and handle partial matches.

Deliverables

Working Python (or Node.js) script/codebase with clear comments.

API endpoints or webhooks for voice and chat input.

Configuration file for CRM mapping (fields & authentication).

Brief documentation (1–2 pages) for testing and maintenance.

Freelancer Qualifications (Preferred)

Proven experience with Claude API (Anthropic) and prompt engineering.

Past work integrating voice (STT) and chat into a single workflow.

Familiarity with at least one CRM’s API (list which you know).

Bonus: Twilio, Make.com, or n8n automation experience.