Back to Projects
AI-Powered Call Transcript Insight Extraction & Sheets Integration
AI workflow that digests 20-minute calls, summarizes the themes, and syncs action items to Google Sheets.
Project Overview
I engineered a Python + Flask backend with a lightweight UI that ingests lengthy support or sales calls, runs them through custom OpenAI prompt chains, and returns structured insights within minutes. The system cleans raw transcripts, extracts topics, action items, sentiment, and customer feedback, then pipes everything automatically into Google Sheets so ops teams can filter, share, or trigger follow-up workflows without touching the raw audio.
Key Features
Transcript preprocessing pipeline that chunks 15–20 minute calls for accurate LLM analysis.
OpenAI-powered extraction of key topics, action items, concerns, suggestions, and sentiment.
Automated Sheets integration so every call lands as a structured row for dashboards or alerts.
Configurable tagging so teams can adapt the model to different call types or verticals.
Skills: Python, Flask, OpenAI API, AI Data Analytics, AI Content Detection.