It’s been a whirlwind couple of days, jumping between a bunch of different tasks. I’ve been trying to standardize some internal processes, providing quick fixes for manual work, and pushing my main video analysis project forward. It’s a real mix of high-level strategy and getting stuck into the technical details.
Here’s a snapshot of what’s been on my plate:
-
The best solution isn’t always AI. I was asked to help with a process where an employee spends 3-10 hours manually comparing printed invoices against our inventory in Excel. My first thought was some complex OCR solution, but after a quick look, I realized the answer was much simpler. We just need to get the invoice as an Excel file from the supplier and use
XLOOKUP. A five-minute fix that saves hours of work, no AI required. - Progress on the video analysis project. We had a good meeting with senior management and got the green light to move forward. We’ve purchased three GoPros to start documenting our assembly lines. The goal is to automatically calculate standard times for tasks, which is a huge need right now.
- Hitting some technical roadblocks. Now for the hard part. I need to figure out the pipeline for the video project. How do I handle super long video files? Should the processing happen locally or in the cloud? Where do we store all this data? And I’m worried that if a browser tab crashes mid-analysis, the results will be lost. Lots to solve here.
- Trying to add value with process flows. I’ve been using Claude to create standardized, visual process flows for our assembly steps. The feedback has been positive, but I’m still trying to figure out how to measure the real-world value this brings.
It’s a good reminder that this job isn’t always about building complex AI systems. Sometimes the biggest impact comes from identifying a simple, non-technical solution that saves people from tedious work. These small, practical wins feel just as important as cracking the big technical challenges.
Next: Start mapping out the video analysis pipeline, focusing on how to handle long files and where to store the data.
Tags: process automation, AI, Excel, video analysis, problem solving, lean manufacturing
Top comments (0)