📌Context:

• Industry: Call Center / Customer Support / Sales
• Process: Processing recorded customer phone calls, transcribing audio, analyzing sentiment, and logging key call metrics into a database.

📌Tech Stack: n8n, AssemblyAI API, OpenAI API (LLM), Google Sheets, Google Drive.

📌Problem:

• Manual Effort: Managers manually listened to full call recordings to evaluate call quality and sentiment.
• Lack of Data: No structured tracking of call duration, cost, or word count metrics in real-time.
• Human Error: Inconsistent sentiment classification and slow feedback loops.

📌Solution:

An automated end-to-end n8n workflow that:

1. Automatically retrieves audio recordings of calls.
2. Transcribes audio into text with timestamps using AssemblyAI API.
3. Analyzes transcript sentiment, key topics, and summary using OpenAI API (LLM).
4. Calculates metrics (duration, word count, transcription cost).
5. Updates Google Sheets with structured results and status Done.
6. Deploys a specialized AI Agent that queries the database, extracts overall statistical insights, identifies call trends, and generates summary performance reports.
7. Handles execution errors via a dedicated Error Workflow.

📌Process:

• Trigger & Fetch: Reading call data from Google Sheets.
• Audio Transcription: Sending audio URLs to AssemblyAI for high-accuracy text conversion.