Adam, the definitive test for your memory question is to call up the Resource Monitor to observe page-fault rates. Despite having all of those things in the process list, it is possible that you could be totally fine. If your system has found an equilibrium mix of resident and swapped-out pages, then you should see only minimal page-fault actions.
The only way to tell whether you have a problem is to set up a loaded but otherwise quiescent system. I.e. that task list you showed us would be a decent system load for the test I'm considering. During the test, you want to not be:
* launching a new task - because tasks don't load. They inswap.
* interacting with web pages - because you will be diddling with backing store, which is a swapping activity.
* running some Access form with a short-interval timer in the background - because you will be forcing heavier Windows Task (internal) Scheduler activity.
* running an anti-virus scan or update. (Not saying "disable A/V" - just saying don't do anything active with it to trigger it.)
* experiencing machine hardware failures at the time (let's hope this one is easy).
Your page fault rate in a quiescent system should be close to zero. It will jump when you launch a new task, but should settle down as the system finds a new equilibrium.
Because of system "round-robin" actions relating to "fairness in scheduling" concepts, you will see occasional fault activity, kind of "bursty" operations in the "every 20-40 second" range. This is normal and the sudden spike should drop off very quickly. In a quiet system, even the low-priority processes can get inswapped if they haven't been given CPU access for a while.
If you have hundreds of page faults per second in a system where you weren't actually interacting with anything, then you have started to "thrash" and in that case, YES you need more memory. It is a complex bit of math to decide how many page faults per second is "too many" but if you ever reach a sustained rate in the hundreds per second, you are beating your system disk to death.
By the way, if you are in Resource Monitor for the page fault test, you should ALSO be able to tell whether any of those Chrome and FireFox sessions are network-active, because that is ALSO a display on the RM. On my Win10 version of RM, network usage is two frames above page-fault rate.
I'm betting that no more than one or two out of those browser sessions actually has that much network activity. If ALL of them exhibit network activity, you need to stay away from so many active pages, 'cause THAT will totally hose your "quiescent system" tests and will make your Win internal task scheduler beat the system to death, maybe literally by exercising your disk more than it would like. Trust me, system-disk crashes are no fun at all. I know from recent experience.