Search for what it costs to automate a process and you will find the same thing everywhere: ranges from €1,500 to €60,000, a list of factors that move the price, and a closing line that sums it all up. The exact figure comes out of a prior audit. That is true. And it is no use at all when what you need is to know whether it is worth it before booking a call with anyone.
Why everyone answers «it depends»
Because it is true, and because it suits them. It is true: the price depends on your systems, your volumes and how many odd cases have to be handled. And it suits them: a range from €1,500 to €15,000 commits nobody, and forces you to ask for a quote. The problem is that the range is so wide it does not help you decide. Between €1,500 and €15,000 there is a factor of ten.
The formula we use, published
We price the build at 12 % of the annual cost of the task being automated. If a task costs you €20,000 a year, the project comes to around €2,400. If it costs €4,000, around €500. The price is tied to what the task costs you today, not to how complex it looks to us. Maintenance is 20 % of that build cost, spread across the year.
The three numbers you need
To work out what the task costs today you only need three figures, and you have all of them. How many times a day it is done. How long it takes each time. And what the hour costs for whoever does it, employer contributions included. Multiply the three by 240 working days and you have the annual cost. No spreadsheets, no odd formulas.
A worked example, with the numbers on show
A company keys in supplier delivery notes twice a day, twenty minutes each time, with someone costing €25 an hour fully loaded. That is 160 hours a year and €4,000. The build comes to around €480 and maintenance to around €96 a year. Now for the other half of the calculation: how much of those €4,000 automation actually recovers.
You almost never recover 100 %
This is where most payback calculations go wrong. No automation removes the whole task: exceptions, checks and cases that do not fit all remain. If the data sits in a business system and the cases resemble one another, you can recover 30 to 50 %. If it sits on paper and every case is different, you drop to 10 or 15 %. That difference decides the entire project.
When the answer is no
We only propose a project when it pays back in under 24 months. Below that threshold there is a case; above it, there is not. In practice that rules out tasks of a few minutes a day, those whose rules change every quarter, and those whose data sits on paper without having been sorted out first. Not because they are impossible: because the cost of keeping them running eats the saving.
You can run these numbers yourself with a calculator, or let a free diagnostic do it: four questions, two minutes, and it returns the annual cost of your task, how much of it is recoverable, the minimum technology required and how many months it takes to pay back. No appointment, no talking to anyone. And if it turns out not to be worth it, it tells you that too.