A student turned raw lesson transcripts into tidy Notion summaries

Project Description: This project automates the creation of detailed lesson summary pages in Notion for each session of the Mindvalley AI Mastery Advanced Course. It leverages Make.com and OpenAI's GPT models to process lesson transcripts, summarize key insights, and populate them into a structured Notion database. The system streamlines content organization and enhances the learning experience by turning raw session data into actionable, easy-to-navigate summaries. This solution was developed to address the challenge of manually organizing large volumes of course material. Despite initial difficulties with Make.com's Notion integration, I persisted and successfully implemented a functional workflow, significantly improving efficiency for myself and my peers in the course. Achievements: 1. Automated Learning Process: - Created a scalable system that automatically processes lesson content and populates it in Notion. - Reduced manual effort by automating repetitive tasks like transcription summarization and content indexing. 1. AI-Powered Summaries: - Utilized OpenAI's GPT models to generate concise, high-quality summaries of lesson content. - Structured summaries into predefined sections, making it easier to review and apply the learnings. 1. Overcoming Technical Challenges: - Successfully navigated the technical hurdles of integrating Make.com with Notion, a feature that was not fully supported. - Collaborated with peers and mentors to refine the automation process. 1. Organization:  - Ensures outputs are systematically stored in tools like Google Sheets, Notion, or Google Drive. - 1. Scalability and Reusability: - The workflow can be adapted for any course or learning material, making it a scalable solution for educational institutions, businesses, and personal development programs. Tools Used: - Make.com: For workflow automation, document processing, and integration with Notion. - Notion: To store and organize the lesson summaries. - OpenAI GPT: For generating summaries from lesson transcripts. - Google Drive: For storing raw lesson transcripts. - PDF.co: To convert PDF files into text format for processing. Methodology: The automation workflow follows these key steps: 1. Google Drive - Watch Files in a Folder: - Description: This module continuously monitors a designated folder in Google Drive for new files, specifically PDF lesson transcripts from the Mindvalley AI Mastery course. Whenever a new file is detected, the workflow is triggered to begin processing. This ensures that no new lesson material is missed, streamlining the intake process. - Function: Monitors a specific folder in Google Drive for new files. - Trigger: When a new file is uploaded, the workflow starts. 1. Google Drive - Download a File: - Description: Once a new PDF file is identified, this module downloads the file from Google Drive to make it accessible for further automation steps. This ensures that the document is available locally within the Make.com workflow for efficient processing. - Function: Downloads the detected file (likely a PDF) from Google Drive to enable processing. - Output: The file is ready for further processing in subsequent steps. 1. PDF.co - Convert File to PDF: - Description: This module extracts the text content from the downloaded PDF file. The extracted text serves as the raw material for the subsequent processing steps. PDF.co ensures that the conversion is accurate, preserving the integrity of the lesson content. - Function: Extracts text content from the downloaded PDF. - Output: Plain text from the PDF file. 1. OpenAI - Create a Completion (Prompt): - Description: The extracted text from the PDF often includes timestamps and unnecessary identifiers that can clutter the summary. This module uses OpenAI's GPT model to clean up the text by removing these irrelevant elements. This step was crucial to address an error that occurred when processing larger files, where unnecessary details would cause confusion and hinder the summarization process. - Function: Processes the extracted text using OpenAI to:  i. Reduce the size of the PDF file by removing the timestamp and number identifications. - - Input: The plain text from the previous step. - Output: A shorter size of the content  1. Google Drive - Create a File from Text: - Description: After refining the text, this module creates a new text file in Google Drive. This refined text file acts as a backup and a clean version of the original transcript, ensuring that the processed content is stored for future reference or manual review. - Function: Saves the refined OpenAI-generated output as a new text file back in Google Drive. - Purpose: Keeps a copy of the content for manual review or sharing. 1. OpenAI - Create a Completion (Prompt): - Description: The cleaned text is then sent back to OpenAI's GPT model with a specific prompt to generate a structured summary. The output is divided into sections such as Key Points, Comprehensive Overview, Action Items, and Resources/Tools. These structured summaries make the content more digestible and actionable for learners. - Function: Processes the extracted text using OpenAI to: i. Summarize the content of the TXT file. - - Input: The plain text from the previous step. - Output: A summarized version of the content  1. JSON - Parse JSON: - Description: The AI-generated summary needs to be formatted for insertion into Notion. This module converts the structured text output into a JSON format, which is a machine-readable structure that can be easily processed by Notion. The JSON format ensures that each piece of information (e.g., Lesson Title, Summary, Key Insights) is correctly mapped to its corresponding field in the Notion database. - Function: Structures the AI-generated output into a JSON format. i. Example: Breaking down the content into predefined fields like "Title," "Summary," or "Key Points." - - Purpose: Prepares data in a structured format for uploading to Notion. 1. Notion - Create a Database Item: - Description: This module takes the structured data from the JSON output and creates a new entry in the designated Notion database. The entry includes predefined fields such as Lesson Title, Key Insights, Summary, and Action Items. This step ensures that the summarized content is systematically organized and easily accessible within Notion. - Function: Adds the structured data from JSON into a Notion database. i. Example Fields: 1. Lesson Title 2. Summary 3. Key Insights 4. Action Items - Output: Creates a new row in the Notion database for tracking lessons or other insights 1. Google Drive - Move a File/Folder: - Description: To prevent duplicate processing and keep the workflow organized, the original PDF file is moved from the "New Files" folder to a "Processed Files" folder in Google Drive. This archiving step maintains a clear record of which files have been processed and ensures that the workflow only processes new files. - Function: Moves the original file from the monitored folder to another location (e.g., "Processed Files"). - Purpose: Keeps the workflow organized and prevents duplicate processing. Learnings: 1. Technical Problem-Solving: - Overcoming the limitations of Make.com's Notion integration required creative problem-solving and persistence. Initially, there was no direct integration available, which necessitated crafting a workaround by leveraging APIs and JSON formatting. It also required troubleshooting errors related to data formatting and structure, especially when dealing with larger transcripts. By iterating through different approaches and testing multiple solutions, I was able to achieve a seamless integration that now functions reliably. - I learned the importance of resilience and adaptability in dealing with technical challenges. Each hurdle was an opportunity to deepen my understanding of automation tools and their capabilities. Through this experience, I developed stronger problem-solving skills that can be applied to future projects. 1. Efficiency Gains: - The automation significantly reduced the time spent on manual tasks, such as downloading transcripts, summarizing content, and organizing data in Notion. Prior to this system, managing course content was time-consuming and prone to errors. With this workflow in place, I can now process and organize lesson content in minutes instead of hours. - This efficiency gains not only benefited me but also impacted my peers. By sharing the automated process with them, I helped others save time and focus more on learning and applying the course material. The streamlined process ensures consistency and accuracy across all lesson summaries, improving the overall learning experience. 1. Collaboration and Iteration: - The project evolved through collaborative efforts and feedback from my peers and mentors. Initially, the automation faced several issues, such as handling large files and structuring the output in a user-friendly manner. Through regular discussions and feedback sessions, I refined the workflow to address these challenges. - I also learned the value of iteration in project development. Each version of the workflow brought new improvements and optimizations. The iterative process helped me stay open to new ideas and continuously improve the system's functionality. This mindset of continuous improvement is crucial for any automation project. 1. Scalability Potential: - One of the key takeaways from this project is its scalability potential. The workflow can be easily adapted to other educational contexts, such as online courses, workshops, or corporate training programs. By simply adjusting the prompts and database structure, the automation can accommodate different types of content and learning objectives. - Beyond educational use, this automation could be applied to other domains, such as meeting transcription, client onboarding processes, or content creation workflows. The flexibility and adaptability of the system make it a valuable tool for a wide range of use cases.