"AI automation" gets thrown around a lot, but for most businesses it comes down to a simple question: which repetitive tasks can a system handle so your team doesn't have to?

You don't need a machine-learning team or a big budget to start. With a workflow tool like n8n and access to an LLM, you can automate meaningful work in days, not months. Here's how I approach it.

Start with the boring, repetitive tasks

The best first automations are the ones nobody enjoys doing:

  • Copying data between two systems that don't talk to each other
  • Triaging and tagging inbound emails or messages
  • Drafting first-pass replies, summaries, or reports
  • Chasing follow-ups on a schedule

These are high-volume, low-judgment tasks — exactly where automation pays off fastest and safest.

Why n8n plus an LLM is a great combination

n8n handles the "plumbing": triggers, connecting to your apps, moving and transforming data, and running on a schedule. It's the reliable, deterministic backbone.

The LLM handles the parts that used to require a human's judgment — reading unstructured text, classifying it, summarizing it, or drafting a response. You call the model as one step inside the workflow, with a carefully written prompt and clear guardrails.

Together, you get automation that's both reliable (n8n) and smart (the LLM) — without handing the whole process over to a black box.

Design for failure from day one

The difference between a demo and a production automation is how it handles things going wrong. Every workflow I build includes:

  • Retries for temporary failures (an API being briefly unavailable)
  • Error branches that alert a human instead of failing silently
  • Logging so you can see exactly what happened and when

An automation you can't observe is an automation you can't trust.

Keep a human in the loop — at first

For anything customer-facing, start with the AI drafting and a human approving. This builds trust, catches edge cases, and gives you real examples to refine your prompts. Once you've seen it perform consistently, you can move the safest steps to full automation.

A realistic first project

A great starter project: automatically read inbound leads or support messages, classify them, summarize the key details, and route them to the right person with a suggested reply. It's genuinely useful, low-risk, and touches every core concept — triggers, data handling, an LLM step, and human oversight.

The takeaway

AI automation isn't about replacing your team — it's about removing the repetitive work that drains their time. Start small, target the boring tasks, build in safety, and expand from there.

Curious what this could look like for your business? Let's talk.