Focus stacking needs to be much more accurate for professional requirements.

Photoshop’s built in-focus stacking is currently a severe compromise for product and macro/micro photographers. It has by far the most convenient workflow in terms of editing, but also by far the worst precision among competitors. Surely Adobe would have all the resources and capability to outshine Helicon Focus and Zerene Stacker if you put your mind to it. Sadly though, focus stacking hasn’t seen any improvement in Photoshop since its inception in 2008 I believe and thus is incredibly inaccurate for complex objects, requiring a lot of manual corrections to the masks which is very time consuming. Zerene Stacker does a much (much, much...) better job of stacking complex photos, but they have their own serious drawbacks (lack of RAW support and many editing capabilities, not to mention the terribly outdated UI). Here’s a list overview of the current focus stacking situation:

• Focus stacking in Photoshop:
  * CON: the worst quality among serious competitors, inaccurate masks
  * PRO: best file support including RAW, no conversions necessary
  * PRO: all the editing capabilities are included in the workflow by default
  * PRO: comes built-in and paid-for within the subscription
  * PRO: better (or at least familiar) UI
• Focus stacking with competitors (Helicon, Zerene)
  * PRO: better quality, Zerene Stacker offers the most accurate stacks by a long shot
  * CON: limited file support, requiring multiple saves to convert between formats
  * CON: limited editing capabilities within software
  * CON: only available as a separate purchase
  * CON: outdated UI, inconvenient to use on modern systems
  * CON: moving between different programs wastes time

By conclusion, only by making Photoshop’s focus stacking engine better, you’d come up on top of the product photography and macro/micro photography game because you already have the upper hand when it comes to the rest of the workflow. For inspiration, certainly check how Zerene does it to set a benchmark for the expected quality.