Depends on the workflow and the business. Previous job: mostly devs and data analysts, some marketing folks, HR and finance and the IT people.
Upgrading the old trashy wifi APs (which were surprisingly running custom dd-wrt) with unifi (from 2 to 5 APs) for better bandwidth and removing deadspots made HR, finance, marketing and management happier. Some people in dev started using wifi as well. I don’t like unifi, but I don’t regret the decision.
For the backend rack stuff, we were on gigabit for the longest time. We actually saw performance degradation on our backups, from moving from 2x HPE 48 gigabit ports (10G backbone between the 2) with balance-alb over to cisco (also 48 gigabit ports) with LACP and the cisco swtich stacking thing. I personally was against the side-grade and wanted to see 10G upgrade on all servers, but that never happened while I was there (I think it eventually happened 2 years after I left).
Despite all servers running on gigabit (4 to 6 ports + IPMI), the biggest advantage to the workflow was having VMs balanced across hypervisors and storage and decommissioning some old servers (ddr2 era core2 intel motherboard servers - we took them down around 2019, those servers were probably more than 10 years in service and a huge liability).
Unless I had no other choice, I wouldn’t be using gigabit these days, except for specific things. But if you’re on a shoestring budget, you can get away with it by doing a ring-network between your servers on a high-bandwidth port (with OSPF) and use gigabit to access the resources. Wendell showed something like this on mini-PCs with thunderbolt networking, but if you can do it with 10G ports, on proper servers, it’s better (apalard did something like that as well).
Heck, if you’re very desperate and the business needs aren’t that big, run your services on a single server, have your data backed up, maybe even replicated (in addition to backups) and configure a workflow to quickly deploy and restore your stuff (it could be in the cloud, if you’re already using something like S3 for your off-site backups or it could be on another weaker server you keep around as a spare). If availability is important, replication is key (for a bit more costs). If that’s not as important and a few hours of downtime isn’t a big deal, just have good (and tested) backups.
As for the brands, I wouldn’t look too deep into it. I just go with mikrotik, cuz it’s cheap and pretty good bang-for-the-buck. But don’t run any proprietary router. My own experience tells me to stay away from pfsense as well. If you’re not invested enough to run an openbsd, freebsd or linux router directly, I hear people liking opnsense and openwrt as firewalls (I’d stick with CLI in all honesty, some things that appliances do to make the user’s life “easier” kinda makes the sysadmin’s life harder IMO).