This project analyzes sales call transcripts to extract critical insights like customer sentiment, unresolved objections, and salesperson performance. It processes various input formats such as audio, video, DOCX, PDF, and YouTube links, offering comprehensive insights to sales teams to help improve performance.
Analyzing multiple formats of sales calls and extracting meaningful insights in real time was a challenge. Ensuring accuracy in sentiment analysis and extracting the right key points from conversations was also complex.
The project uses Whisper OpenAI for transcription, Langchain for process chaining, and GPT models for insight generation. FFmpeg was employed for audio and video processing to ensure smooth transcription and analysis.
The system provides valuable insights for sales teams, improving their understanding of customer interactions and enabling better decision-making for closing deals.