I found some XPaths chosen by Ui.Vision from recording don’t work well. It’s because many web pages use dynamic HTML. I have to research and change these XPaths to non-dynamic XPath every time. I tried to use OCR to solve this issue. But OCR sometimes can see the defined word sometimes can’t see it (yes, the same word). It feels that manually inputting a non-dynamic XPath is a more reliable solution compared with OCR.
Is there a better or workaround solution for this issue?