Today I mapped out an interesting automation project I’m about to embark on. The challenge is something many brands face but rarely solve properly. A company has thousands of user-generated videos stored in Google Drive. The content shows dogs wearing their products, but the entire library is chaotic. No tags, no metadata, no searchable structure — just folders filled with videos collected over several years. Finding the right clip for marketing or social media becomes almost impossible. So I designed an AI-powered automation system using Make that will automatically organize the entire video library. The system will work in **three stages. Stage 1 – Video Library Indexing The first automation will scan the entire Google Drive library, including nested folders. Every video file will be logged into Google Sheets with key details like file name, file ID, link, and size. This instantly converts thousands of scattered files into a structured video database. **Stage 2 – AI Tagging Engine** Next comes the intelligence layer. The automation will read rows in Google Sheets where tagging is missing, download the video, and send it to Gemini for analysis. The AI will return structured JSON metadata such as: • Dog breed • Product type • Product color • Content type (UGC, review, demo, etc.) • Setting (indoors, outdoors, park, home) • Season • Usability rating for marketing • A short content description The automation will then parse the JSON and automatically write all the metadata back into the spreadsheet. Stage 3 – Continuous Automation Finally, a monitoring workflow will watch for new uploads in Google Drive. Any new video added to the folder will automatically pass through the same pipeline — analyzed, tagged, and added to the structured database. The result will be a fully searchable AI-organized video library. Instead of scrolling endlessly through folders, the brand will be able to filter content instantly by: • Dog breed • Product type • Scene or setting

Last week I rebuilt a complex Alteryx workflow inside n8n for a client. The goal was simple Stop running manual data routines across multiple tools and turn everything into one automated pipeline. Before this setup the client had to run the workflow manually every time they needed to process new data. It involved multiple steps data cleaning transformations and sending the results to different systems. So I replicated the entire logic inside n8n. Now the system automatically pulls data processes it cleans it transforms it and sends the final output to the correct tools without anyone touching it. What the automation handles now Replicates the full Alteryx workflow inside n8n Automatically processes incoming data Connects multiple APIs and data sources Cleans and transforms datasets automatically Runs on schedule or real time triggers Sends processed data to the correct systems instantly The interesting part was translating the original Alteryx logic into nodes conditions and data transformations inside n8n while making sure everything stayed reliable and scalable. What used to be a manual routine is now a fully automated data pipeline running quietly in the background.