Free tool
Not the one that annoys you. The one that costs you.
Most first automation projects pick the most irritating job rather than the most expensive one, and that is the single most common way they disappoint everybody. List yours and see them in order.
Score your jobs
In this order
3 jobs scored.
- 1Start here
Typing job sheets into the accounts system
- Eats 125 hours a year, which is the main reason it is here.
- Runs often enough that the process is already well worn.
- Touches 3 different systems, so somebody is copying between them.
- Same rules every time, which is exactly what a machine is good at.
- 2Worth doing
Building the Monday report
- About 72 hours a year.
- Touches 4 different systems, so somebody is copying between them.
- Same rules every time, which is exactly what a machine is good at.
- 3Later
Chasing clients for missing documents
- About 48 hours a year.
- Goes wrong often, and fixing mistakes costs more than the job does.
The audit form arrives already carrying your top three, so you will not be asked to describe them again.
What it is weighing
Time, mostly
Hours a year is half the score. Everything else is a modifier on top of it, because a job that eats a month of somebody's year is worth doing whatever else is true of it.
Rhythm, separately
A daily job beats a twice-yearly one of the same total size. It is practised, predictable, and already documented by the fact that somebody does it constantly.
How many systems
Every extra piece of software is another place somebody is retyping what they typed a minute ago. That gap is where most of the value is.
Judgement, subtracted
The one thing that pulls a score down. Automation follows rules well and handles "it depends" badly, and pretending otherwise is how a project ends up half-built.
The usual questions
What is it scoring?
Four things. How many hours a year the job eats, how often it runs, how many separate systems it touches, and how often it goes wrong. Then it subtracts for judgement, because automation is good at rules and bad at 'it depends'.
Why did it say a job is not worth automating?
Because some are not. Anything running four times a year or under about twelve hours a year gets refused however well it scores otherwise. Building something to run four times a year costs more than doing it by hand, and that is the same answer you would get from us on a call.
Why does needing judgement lower the score?
Because a job where every case is different is the hard kind. It can usually still be done, but the honest version automates the routine seventy per cent and leaves the rest with a person, and that is a smaller win than it first looks.
Is the top answer always the right place to start?
Usually, but not always. A job that scores second and touches a system you are about to replace is a better first project than one that scores first and depends on software leaving in March. That is the kind of thing the audit catches and a score cannot.