Improve PDF Handling

Yesterday I wanted to compress a PDF to a smaller size. So using substage I queried this (using chatgpt 4.1 model) and to my delight it worked and converted to indeed a smaller size. Then I had a second file I wanted to do this with. And to my surprise, it came back saying it didn’t know how to do it. Used the same prompt. Just wondering if the training built into the app can help handle PDF actions a bit better. Not sure why but even after trying it with several different prompts, it was never able to do it. But I know it was able to understand and do it the first time.