n8n, Make or a custom AI agent: what each one is and which suits you

n8n is a tool for connecting software and automating tasks with visual workflows (when this happens, do that), which you can install on your own server or use in its cloud. Make and Zapier do something similar, but only as a cloud service and with less technical freedom. A custom AI agent is something else: it doesn't follow a fixed workflow, but understands each case (an email, a PDF, an order written any old way) and decides the next step within the limits you set. In short: if your data always arrives in the same format, n8n or Make will do; if every case is different and you want someone to handle it for you, you need an agent.

What is n8n?

n8n is a workflow automation platform. You build a workflow by linking "nodes" (Gmail, a spreadsheet, your CRM, an API request…) and n8n runs it every time something happens. Its strengths:

• It can be installed on your own server, so your data stays in-house and you don't pay per execution.
• It's very flexible: you can add code when the visual workflow isn't enough.
• It includes AI nodes, so a workflow can call a language model to summarise, classify or draft text.

The downside: someone has to set it up, maintain the server and fix it when something changes.

n8n vs Make vs Zapier

• Zapier: the easiest to get started with and the one with the most ready-made integrations. It charges per task run, so it gets expensive at high volume.
• Make: more visual and more powerful than Zapier for workflows with branches, also cloud-based and pay-as-you-go.
• n8n: the most flexible and the only one of the three you can host yourself. It needs a bit more technical knowledge.

All three are tools for building workflows: you decide every step in advance.

Where does workflow automation fall short?

When cases don't always arrive in the same way. An order that sometimes comes in the body of the email, sometimes as a PDF and sometimes as a photo. A customer who asks three things in one message. An invoice that has to be classified depending on what it's for. With fixed workflows you end up with dozens of branches and exceptions that someone has to maintain. That's where an agent comes in.

What does a custom AI agent add?

An agent is given a job ("take the orders that arrive by email all the way into the ERP") and does it case by case: it understands what comes in, decides what to do, uses your software and checks with you on anything you've marked as sensitive. What's more:

• It works on its own, on its own schedule: it doesn't wait to be triggered.
• It asks your permission for important things and alerts you on your phone.
• It adapts to the way you work, instead of you adapting to a template.

We explain it in depth in what is an AI agent.

Has n8n become obsolete with AI agents?

No. They're different pieces that complement each other: an agent can use n8n workflows underneath for mechanical tasks (moving a file, syncing two programs) and save the intelligence for what needs judgement. The useful question isn't which tool is better, but who makes the decision in each case: a fixed workflow or an agent with limits.

Which one suits you?

• Zapier or Make if you want to build it yourself, volume is low and your data always arrives in the same format.
• n8n if you have someone technical, want your data on your own server or have high volume.
• A custom agent if cases vary, you want the work to get done on its own and only what you need to sign off to reach your phone.

At balderas.studio we build custom agents, from €990 to set up and €149/month, and if a simple workflow is enough for your case we'll tell you: sometimes the best solution doesn't involve AI. More detail in how much AI automation costs or book a 30-minute call.