n8n + AI for DevOps processes

It’s 2025, and AI is everywhere — in articles, presentations, and even during coffee breaks at meetups. But when you try to bring it into your team’s actual workflows, it suddenly turns out to be expensive, time-consuming, and full of cross-department communication hurdles.

We have a strong and very busy Data Science team — they’re working on big, business-critical projects: models, RAG systems, bots for flagship products. So coming to them with a request like “can you automate my Telegram a bit?” feels like using a cannon to shoot a sparrow.

What we really needed was a tool to build simple MVPs ourselves — quickly, without distracting the heavy artillery. So we took on the challenge and decided: let’s simplify and automate. In a way that lets you build something fast, test it, throw it away, or hand it over to a proper team for polishing.

That’s how we discovered n8n — a visual automation engine that’s open-source and surprisingly usable. Why not Zapier, Make, or IFTTT? Because they’re SaaS, and we needed on-prem. Something we could plug into our infrastructure, keep sensitive data in-house — and still have a UI that even managers can figure out.

In this talk, I’ll share our experience with n8n: how it came into our lives, the pitfalls we hit, and where it’s already working today. We’ll talk about how we build self-service automations, the early-stage screw-ups, and why even HR is now part of the n8n workflow.

Promise: no marketing fluff — just real-world experience, memes, field notes, and practical results.

Disclaimer: We only installed and started rolling out n8n in May 2025. So some use cases are still in the pipeline, and the tool is currently being battle-tested inside the company. If it turns into a failure story — the talk will just pivot to that instead.