Process automation with n8n and Make: when does it pay off?

At a glance

• Process automation pays off when a workflow comes up often, follows fixed rules and takes up working time today.
• The calculation is short: hours per month times your hourly rate gives you what the workflow costs today.

How do you work out whether a workflow is worth automating?

Take one workflow and measure how many hours it takes each month. Multiply that by your hourly rate, including overheads. HAGEL IT uses this worked example on its website (according to my market research of 1 October 2026): an employee who spends 10 hours a week on data entry costs €1,400 a month at an hourly rate of €35. Put in your own figures. The result shows how much room you have for an automation.

Which workflows are worth automating?

Typical candidates are workflows with fixed steps that run across several programs:

Workflow	What runs automatically	Where a person stays in charge
Incoming invoices	email arrives, the document is read, fields are extracted, matched against the order, passed on for approval, booked in the accounting system	approval and exceptions
Writing quotes	an enquiry arrives, your text blocks and prices produce a draft, a reminder goes out if there is no reply	checking and sending the quote
Onboarding	contract, logins, equipment list and checklist are created, reminders go to the manager	conversations with the new employee
Handling enquiries	the enquiry lands in the CRM, confirmation of receipt, suggested appointment, follow-up	the advice call

The pattern is always the same: the manual workflow becomes an automatic one, and you decide where a person makes the call.

Where does AI come in?

An automation follows fixed rules. You bring in AI when a step needs to understand, sort or write text, for example with incoming emails or when extracting data from documents. Often you combine both. Decisions with consequences stay with a person.

When does process automation not pay off?

Usually not when a workflow rarely happens, when it changes with every case, or when nobody can describe exactly how it runs today. Assess every workflow in three categories: can it be automated now, after some preparation, or better not at all? Leaving a workflow as it is can be a good result.

Do you need n8n or Make?

Both tools connect programs into workflows. Which one fits depends on your existing systems, your data protection requirements and the licence fees, not on the tool itself. According to my market research of 1 October 2026, licence fees for Make, n8n Cloud, AI APIs and text recognition are typically €10 to €80 per month. You pay them to the provider; I add no mark-up.

When is an internal tool better than an automation?

When the workflow lives not in programs but in Excel spreadsheets sent round by email, a dedicated tool built on your existing systems and data often helps, for example an approval screen or a customer portal. The automation runs behind it.

What does process automation cost?

According to my market research of 1 October 2026, quotes for automated workflows range from €1,500 to €9,900: small automations from €1,500, automation and RPA from €2,500 and from €2,800, a four-week package up to a running workflow for €9,900. HAGEL IT quotes €5,000 to €30,000 for bespoke solutions. An upfront analysis costs €950 to €990 on the market. Not every price says whether VAT is included.

How does a project work?

1. First I map the workflow with the people who do it today: who does what, in which program, how often, for how long.
2. I build on your existing systems. You do not replace anything.
3. I test with your real data before the workflow goes live, and we review the results together.
4. Your team is trained on exactly this workflow. Technology, logins and data protection (data processing agreements, data flow, deletion rules) are documented.
5. In the end one number counts: hours saved per month, measured against the starting point.

What happens when the workflow makes a mistake?

An automation that silently fails on a bad document is worse than none. That is why the error case, an alert and a way out for a person belong in the build from the start. After launch, someone should keep an eye on the runs. Funding is possible but not promised; the rules are covered (in German) in Digitalbonus Bayern für KI-Projekte.

Next step: in the AI consultation for €100 (no VAT is charged under §19 UStG), we discuss which of your workflows pays off first: book a call. What I build is described under AI automation.