n8n or Make: which automation tool should you choose?

Automatically sorting emails, generating a quote from a form, sending a consolidated report every Monday: these tasks no longer need a developer. So-called no-code or low-code platforms let you connect your applications and insert AI models along the way. Two names come up again and again: Make and n8n.

Both are excellent. They don’t answer the same constraints.

Make: speed and a huge catalogue

Make, formerly Integromat, is an online service. You build scenarios in a very readable visual interface by chaining modules: “when an email arrives → extract the attachment → save it to a folder → notify the team”.

Strengths:

• quick to learn, including for non-technical people;
• a very large catalogue of ready-made connectors;
• no infrastructure to manage: everything is hosted by the vendor.

Watch out for: billing depends on the number of operations executed. A scenario that runs often on a lot of data can cost more than expected. Your data also passes through the vendor’s servers, which should be assessed under the nFADP depending on its nature.

n8n: control and flexibility

n8n also offers a visual editor, with one major difference: it can be installed on your own infrastructure. Its source code is available under a so-called fair-code licence that allows internal business use. A paid cloud version also exists.

Strengths:

• data can stay with you, on a server in Switzerland for instance;
• when self-hosted, no per-operation billing;
• lots of flexibility: you can drop in code when no-code isn’t enough;
• good tooling for building AI agents that chain several steps.

Watch out for: self-hosting means someone has to install, update and back up the server. The interface is also a little more demanding for beginners.

The deciding criteria

Criterion	Lean Make	Lean n8n
Sensitive data (clients, HR, health)	Depends on assessment	Yes, self-hosted
Team without technical skills	Yes	With support
High execution volumes	Watch the cost	Advantage
Start this week	Yes	Longer if self-hosted
Complex logic, AI agents	Possible	Very comfortable

Where does AI fit in?

Both platforms can call the major models (Claude, GPT, Mistral…) at any step of a flow. That is where automation changes nature: you no longer just move data around, you understand it. Classifying an email by intent, extracting amounts from an invoice, drafting a first reply: these tasks become automatable.

One simple rule applies to both tools: keep a human in the loop for anything that goes to a client or commits the company. The automation prepares, a person approves.

Our recommendation

• You want to test quickly a first use case with no sensitive data: start with Make.
• Your flows touch confidential data, or you expect high volumes: go with n8n, ideally self-hosted in Switzerland.
• When in doubt, the tool matters less than the process: a well-designed flow can be moved from one platform to the other.

We train teams on both tools and build their first automations with them. See the automation service or our no-code training.