AI-Driven Transcript Summarization and Evaluation Workflow

This n8n workflow automates the process of summarizing video transcripts stored in Google Drive and evaluating the quality of the summaries generated by AI language models. It begins with a webhook trigger to receive input, such as a video URL, then downloads the transcript from Google Drive. The transcript is processed and summarized using advanced AI models like GPT-4 and Google Gemini. Subsequently, the workflow evaluates the summarization’s accuracy by comparing the AI’s output against the original transcript, assigning a score reflective of the model’s ability to produce grounded and faithful summaries. The results, including scores and reasons, are stored back in Google Sheets for further analysis. Practical for content creators and researchers, this workflow helps ensure AI-generated summaries are reliable and rooted in source material, improving content quality and AI trustworthiness.