Many years ago I was working as a IT tech trainee, and our company had gotten a contract to upgrade some ~350 local office Unix servers to NT4 for this company. The clients used Windows 3.11, with home directories shared over SMB. Number of users varied between 20-30 to a few hundred.
Clients would still be running win3.11 on the same 486's, but instead a trimmed netboot variant which connected directly to a remote desktop on the NT4 server (Citrix).
The NT4 server would be configured and set up centrally and then shipped, us techies "just" had to copy over the user docs and mail to the new server.
See each user had a home directory and a mail directory, and after copying the data we had to go to each directory and set up the permissions. We've gotten documentation on how to do it, which involved the usual right-click property dialogs, finding the right user in the list etc.
This process was very tedious and labor intensive, not to mention error prone, and was the majority of the several days of on-premise install time.
After a couple of installs my inner programmer took over, and decided there had to be a better way. So I spent a day writing a few batch files that would copy the data, and then iterate over the directories and calling "cacls" directly. At the end I had a small tool which sent an email to an SMS gateway.
After that all we had to do was hook up the new server, start the batch file and go for a 2-4 hour lunch until we got the SMS. Then we could verify the log files from my batch files that everything went ok, possibly fix one or two small issues and go home. On most installs we were done the first day, rather than spending 3-5 days.
I distributed these files to the group and got a lot of praise from the other techies. Then my boss popped by and that's when I learned that in the contract, they would get paid by the number of users per location as they had estimated the work by that. So we would still get paid for 3-5 days of work, while the vast majority would be done in less than one day, and none went more than two days after my scripts were used. So me, a trainee, earned the company a quite significant amount of money by being lazy.
On an side note, this project also taught me the importance of planning.
This project occupied the majority of the available techs in our company during the 3 months of rollout. Servers were prepared at our headquarter and shipped so that techs started install on Mondays, and each Monday for 12 weeks roughly 30 servers got installed. The rollout went flawlessly and on schedule, with only a disk cable coming lose during transport of the very last server causing an issue.
However, the project group had been sitting next to my office so I knew they had spent one and a half years planning and testing. A significant portion of that was the logistics of the rollout itself.