This topic has been discussed many times already. You might
- want to have a look at the documentation (everything under “Automation”, especially the sections on smart rules
- search the forum for “rename date” or something like that
- have a look at Using a single smart rule instead of Hazel (maybe) (which I’m using also for my DKB account statements, btw).
Automation is sensible for things that occur often. In your example case, writing a rule/script to automatically set the name from the content might just be too much work. Banks are not changing their conditions more than one or two times a year, I guess.
Having said that, I’d suggest to think about automating everything that occurs sufficiently regularly. In my case, that’s account statements, recurring telecom and other bills – all arrving once per month. Everything else, I do manually. To take care of every possible subtlety in a smart rule or a script is just too much work.
Also, you should be aware that PDFs as such might not contain a date. Only if they also contain a text layer can you hope for a document date. So a smart rule (or script) should always check if a text layer is available and run OCR on the document if not. A notable example for this missing text layer are invoices from German Telekom: while their fixed line invoices are perfect, mobile ones have to be OCR’d first.