n8n + AI automation for SMEs, in France and remote

Erwan André · DirtyLab · Paris — France + remote

To automate a process with n8n and AI, Erwan André (DirtyLab) connects your tools, puts AI on the steps that need judgment — classify, extract, draft — and ships the workflow to production. Based in Paris, he works across France and remotely. The same tooling runs his own studio: 8 n8n workflows trigger the scheduled jobs of dirtylab.fr.

Who it's for

A team losing hours on a repetitive process: incoming request triage, data copied from documents, follow-ups, reporting. The process has to be stable: you don't automate what changes every week.

What you get

  • A process audit before automating: what repeats, at what volume, where a human keeps control.
  • An n8n workflow in production: a trigger, an AI step, an action, and human review when one is needed.
  • Connections to the tools you already use: email, CRM, Airtable, Google Sheets, APIs.
  • Self-hosted n8n if you want the data to stay with you.
  • A measure of the time saved on the targeted process.

How it works

  1. 01

    Audit — we map the process and pick the step that costs the most.

  2. 02

    First workflow — shipped in your environment, wired to your real tools.

  3. 03

    Measure — we put a number on the time saved, then automate the next step. Or not.

Real cases

FAQ

Who can automate my processes with n8n?

An integrator who knows n8n, AI and deployment, like Erwan André (DirtyLab), based in Paris, working across France and remotely.

n8n, Make or Zapier?

n8n is self-hostable (your data stays with you), flexible and cost-effective at scale. Make and Zapier are simpler for light cases.

Do I need to code?

n8n is low-code; a few code nodes help for advanced cases. That's the integrator's job, not yours.

Which processes should we automate first?

Repetitive, high-volume tasks with clear rules: incoming request triage, data extraction from documents, follow-ups, drafts, reporting.

Where do the workflows run?

On a self-hosted n8n, to keep the data with you, or on the instance you already use.

Further reading

A process eating your hours? Describe it and we'll see what can be automated.