Why the work grows and the hours stop being enough
As long as every case takes the same twenty minutes of a person in front of a screen, the arithmetic is simple, twice the cases want twice the hours, and that is where the question about hiring comes from. But first it is worth looking at where those hours really go, because a good part of them goes on friction and not on work
Take any case and follow it from beginning to end, without stopping at what looks to you like the real work. Count the minutes spent finding the right document, opening the business software and finding the place where you left off, checking again a figure you had already entered, answering the colleague who asks how far along you are. Nobody sees those minutes because they are scattered, but added up they are the part that is holding you back
Separate the work that needs a head from the work that needs only fingers
The split you need runs inside a single case, not between one department and another. On one side there is reading a document, copying a number from a file into the business software, downloading a receipt from a portal, checking that two amounts match, all things that follow a fixed rule and that anyone would do the same way. On the other side there is working out why a customer has asked for something odd, deciding what to do with a case that falls outside the rules, talking to somebody
The first group can be taken out of people's hands, the second cannot, and it would not even be a good idea. The trouble is that almost always your best people spend more time on the first group than on the second, and that is where the time gets thrown away
Do the sum on the hours, with made-up numbers to show the method
You don't need a long analysis, you need an honest multiplication. Take the number of people who do that work, multiply it by the hours they spend on it every day, then by the days worked in a month, then by the hourly cost to the company, meaning what that hour really costs you all in, not the net figure that lands on the payslip
Say, only to show how it is done, that it is two people for two hours a day over twenty days a month, that makes eighty hours. Multiply those hours by your hourly cost and you have the figure for the month, by twelve again and you have the one for the year. These numbers are invented here and yours will be different, but the figure that comes out is almost always higher than the one you would have guessed
Then ask yourself what share of those hours follows a fixed rule. Don't go all the way to a hundred per cent, stay cautious, because part of the human checking stays and has to stay. But even stopping at seventy per cent, that is fifty-six hours a month coming back to you without hiring anybody
Backlogs and peaks behave differently
A backlog is work that piles up because nobody has had the time to do it, and the repetitive part does not have to be done in office hours. It can run in the evening or early in the morning, when the program is free and nobody is using it, so in the morning you find the cases already prepared and the person only looks at the ones that stopped
A peak is a different thing. It is a temporary rise that does not justify a new contract, and if you hire for a peak the cost stays with you once the peak has passed. The right question is whether it hits the repetitive part or the part that needs judgment, because only in the first case can you absorb it without touching your headcount
What to watch to tell whether it is working
The hours saved are the most obvious number, but on their own they say little, because freed hours only matter to you if somebody is using them for something you could not do before. Look as well at how many cases you close in a week, how long it takes from a document arriving to it being posted, and how many errors turn up downstream and cost you a phone call or a credit note
Measure these numbers before you change anything, otherwise in six months you will have the feeling that things are going better without being able to prove it, and a feeling does not hold up when you have to decide whether to carry on
Where it is best to start
Not from the most important process, but from the most boring and most regular one, the one everybody in the office hates and that always goes through the same steps. It is easier to describe, it is easier to tell whether it worked, and if something goes wrong the damage is small
Write the steps down one by one, the way you would explain them to somebody new, odd cases included, because those are what decide whether a thing can really be automated. By the time you reach the end of that list you already know far more than you did before, even if you then decide to leave everything as it is