1 Recognise duplicates when one of them is a clear pdf and the the other has been scanned from paper. (sometimes I get stuff in mail from government even though the same thing has been delivered digitally)
2 Generate sensible filenames based on content.
Not sure about the first one, but the second one is even added as an example smart rule.
And a suggestion: you might want to change the title of this topic. It’s not clear what your topic is about unless we open it and read it as a full sentence.
Thank you. (hard to come up with a better title but I tried)
What about “Can AI in DT4 recognize duplicates and rename files based on content?” 
As a follow-up: can the ai detect generic filenames (scan001.pdf or sdfhjihng-3414.pdf) from “handcrafted” ones only rename the former? Is the smart rule engine flexible enough for that?
Depending on the used AI model it might be possible. But I wouldn’t expect it to be error proof - a disclaimer valid for all kinds of AI usage actually.
As the generic filenames are possibly all based on a certain scheme (or few of theme), it’s probably more reliable, efficient, faster & cheaper to match these filenames via DEVONthink’s smart rule conditions and then to perform the renaming, either via actions, scripts or AI usage.
This batch processing configuration seems to work when e.g. using Gemini 2.0 Flash Thinking but not with Gemini 2.0 Flash:
Thanks for giving a little insight of how this will work.
