Business process automation has no price list — the cost depends on how many processes you cover, how many systems need to be connected and how many exceptions the work itself contains. A small integration between two systems and a full bespoke build differ by multiples. The more important measure, however, is not the absolute sum but the payback: a well-chosen first process typically pays for itself in 3–6 months. The rest of this article looks at exactly that — what makes up the price and how to judge whether a given project is worth it.

What determines the cost of automation?

When we scope a project we look at five things — and they explain almost the entire difference between cheap and expensive projects.

1. The number and complexity of the processes. Automating the sending of a proforma invoice after a confirmed order is one thing; the full cycle from order through warehouse to bookkeeping is another. The more steps and participants, the more work on analysis, development and testing.

2. The systems that need to be connected. Modern software with a well-documented API connects quickly. An old accounting product with no integration interface calls for workarounds — and those cost time.

3. The exceptions. The 80/20 rule applies in full: the standard flow automates easily, but every "except for this client we do it differently" adds logic, tests and maintenance. Sometimes it is cheaper to leave the exception manual.

4. The quality of the data. If the product catalogues are inconsistent and customers have been entered three times under different names, data clean-up comes before automation. It is often underestimated in the budget.

5. The maintenance requirements. Processes change — a bank, a supplier or a tax authority requirement changes with them. Decide who will maintain the solution and how, because that is part of the real price.

What project types are there and how do they differ in scale?

Instead of specific sums — which depend on each individual case — it is more useful to think in four project types, ordered by growing scale.

Connecting two systems. For example, the online shop feeding orders directly into the warehouse software. Scale: days to a few weeks of work. This is the most common entry point.

An automated workflow. Approvals, notifications, reminders — an expense request moves along a chain and everyone can see how far it has got. Scale: a few weeks, including set-up and training the team.

Automating document work. Incoming invoices, contracts, handover protocols — the document is recognised, the data is extracted and enters the system without retyping. The scale depends heavily on how varied the documents are.

A full bespoke solution. When the process is the core of the business and no off-the-shelf product covers it. These are projects measured in months — and it is precisely here that the upfront analysis is most valuable, because a mistake in the brief costs the most. What such a project includes with us is described on the page for our automation service.

No-code or custom — what does the choice mean for the budget?

Off-the-shelf no-code tools have a low entry price: you pay a monthly subscription and start quickly. The bill, however, grows with the number of users and the volume of operations — and after two or three years the subscriptions can overtake the cost of a bespoke build. A custom solution is the opposite: a higher initial investment, but no per-user licences and built exactly around your process. When each option makes more sense is covered in detail in our article "No-code or custom automation: how to choose".

Which costs are missed most often?

Four items are regularly missing from the initial calculation. Licences and subscriptions — with off-the-shelf tools they are a permanent monthly cost that grows with the business. Training and the transition — the team needs time to change its habits, and in the first weeks the old and the new way of working often run in parallel. Maintenance — every change in the connected systems may require an adjustment. And preparing the data — clean-up, unifying catalogues, migration. A good proposal shows all of this upfront rather than discovering it along the way.

How do you judge whether it pays off?

The maths is simpler than it looks. Take one process and answer three questions: how many hours a month does the manual work on it consume; what do those hours cost at the real rate of the people doing them; and what do the errors cost you — a wrong delivery, a missed deadline, data entered twice. Add it up and compare with the price of the project plus a year of maintenance. If the investment comes back in under a year, the process is a good candidate — in our practice the average reduction in costs across automated processes is around 22%, and the typical payback on a well-chosen first project is 3–6 months.

The full method, with a step-by-step example, is described in our article "The ROI of digitalization and automation: how to calculate it".

How do you start without risking a large budget?

Do not start with the biggest process — start with the clearest one. A good pilot candidate has three traits: it repeats often, it has few exceptions and its result is easy to measure. Such a project is small in budget, quick in timeline and convincing as proof for the team and for yourselves. Only after it does it make sense to plan a broader programme. Exactly how to pick the first process is the subject of our article "Business process automation: where to begin".

And one final rule: ask the vendor not for a quote "in the dark" but for a short preliminary analysis. Half a day of conversation over the real process is the difference between a price you can trust and a number that will change three times.

Frequently asked questions