Hi Ben,
I fear that in your case you'll have to generate a new model for the whole area every time and then compare it to the previous model in order to accurately identify and assess changes. It would be quite incredible if PhotoScan knew what has changed before even processing your imagery. PhotoScan is a great piece of software, but cannot do magic
However, PhotoScan allows you to automate/simplify your processing workflow (more details later).
To mitigate/solve some of your problems, I suggest to get more powerful hardware and/or to change processing settings. This should help to accelerate your workflow. For example, if your client just wants a DEM (+contour lines) and/or a orthophotograph, you can choose HEIGHT FIELD during processing which saves already a lot of processing time. Additionally, it might be sufficient to build the dense cloud at MEDIUM quality settings (e.g., many mines don't care about a few cm). Your models should be fit for purpose, that's all
If you fly the same (UAV) survey over and over again to detect changes, you could definitely automate your workflow either by simple
BATCH PROCESSING or more advanced
PYTHON SCRIPTING. The latter gives you more flexibility. I would, for example, use 'fixed' ground control points that don't change and are used in every model. Makes it much easier to align/compare models later.
Depending on your project, PhotoScan's DETECT MARKER function might come in handy as well. You would need to adjust the size of your ground control points to your GSD though (details in other posts of this forum). Photoscan would then assist you in finding markers in all images, which can be time-consuming when done manually.
Regards
SAV