We should also consider that there was a time before personal and home computers. Businesses were ran quite successfully when everything had to be done by hand - writing letters and documentation, balancing checkbooks and setting budgets, keeping registers and inventories etc. The availability of software vs the investment costs vs the size and profit of your company probably decided to which degree you wanted to digitalize it.
A company usually invests in machinery in order to make savings or earnings from it, whether it is in form of money, time or just to make it look better to others. I'm sure there were several "serious businesses" in the early 1980's who quite could not write home an investment on a $2000+ computer, but could afford a $500 computer. Of course the real drawback is that if they locked themselves in with C64 software, migrating to IBM PC later on when the company so required and could afford it, would be troublesome. Possibly that was a factor for some companies that thought of themselves as "serious businesses" to overspend on PC with room for expansion, so they would be less locked in if and when they needed to migrate. Thus I think a lot of early PC's, and likely a good deal of Apple systems, were grossly under utilized.
Electric typewriters were a huge business right before the shift to personal computers.
Anyone remember when corporations had secretarial pools filled with people dedicated to taking dictation and typing for people?
Some companies had people that just retyped documents all day long.
Photocopiers and personal computers caused a drastic change in how businesses did things.
I think it's clear that a lot of 8 bit machines were used for more than games. Most people at least know someone that used an 8 bit for word processing, spreadsheets, managing the checkbook, or for a database. (remember when recipe file programs where a reason to buy a computer?)
This is a bit beyond the scope of the original question, but when it comes to more serious business use, I think word processing with built in mail merge was the first killer business app for personal computers. That would have been 1978 with Wordstar, then came Visicalc and dBase in 1979. The shift to PCs was becoming apparent by 1984 with Lotus 1 2 3, Wordperfect, and Microsoft Word starting to take over. Wordstar sales were falling quarter by quarter in 1984.
In spite of the shift to PCs, the TRS-80 Model IV stayed in Tandy's lineup until 1994.
I think that is later than any other 8 bit in the US except the C64, which was discontinued the same year.
Amstrad's PCW series appears to have lasted a few more years in Europe.
I'm guessing the late exit from the market for the Model IV and PCW was due to businesses still using them.
I'm pretty sure the C64 stuck around that long mostly because of games.
FWIW, Tandy originally intended the Model IV to have a Z800.
The Model IV would have been able to run all Model III, CP/M, and new 16 bit software all at faster speeds, with access to a lot more RAM, and yet at a lower price than PCs.
It would have given Tandy quite an advantage with small businesses in 1983, and the Model V prototype was supposedly faster and had color support.
When Zilog didn't release the Z800 it pretty much screwed up Tandy's plans. It was released as the Z280 in '87(?) but by then Tandy must have thought it was too late.
Funny thing, Zilog delayed the Z800 so they could produce the Z8000.
The Z8000 was to be used in a new Commodore, but Commodore dropped that machine in favor of the Amiga.
Zilog probably lost a lot of sales over that decision. Who knows how many other Z80 based machines would have followed the Model IV's switch to the Z800.