I built a 10-class AI automation curriculum. Here's what 60 people built with it.

Six weeks ago, I sat down and designed a course meant to take someone who had never opened n8n before and get them shipping real, usable automations by the end of it. Not toy examples. Things they could show a client and say "I built this."

Sixty students went through it with ICT Bangla, twice a week, for six weeks straight.

Here's the shape of what I built.

The first stretch was foundations: the n8n canvas, trigger and action nodes, connecting real credentials and APIs, then adding logic and the first LLM into a workflow. By the end of this part, nobody was starting from zero anymore.

From there, the course moved into multi-channel AI agents. Telegram bots that could hold a conversation. Slack-integrated meeting agents. A voice-enabled Telegram agent. LinkedIn content automation. Facebook and WhatsApp customer support agents, including the manual Twilio and WhatsApp Business API setup most tutorials skip entirely.

The last stretch pushed further: a UGC ad generation pipeline using Fal.AI for image-to-video ads, and using Claude to generate entire n8n workflow JSON directly, letting AI write the automation instead of just living inside it. The course closed with a class on turning all of it into a portfolio and landing the first freelance client.

That's the system. Here's the proof it worked.

What they actually built

These aren't sandbox exercises. Several of these are already sitting in personal portfolios, built by people who had never touched n8n six weeks earlier.

What they said after the last class

One student wrote that the course changed how she thought about learning tech entirely, and that the teaching style made complex topics feel manageable from the very first class.

Another, joining from the UAE, said the practical, project-based approach gave him the confidence to consider freelancing in AI automation, something he didn't think was possible for him a few months ago.

A few asked if their friends and family could join the next round. One called it the best learning environment he'd been part of, and specifically praised the patience behind every explanation of AI automation, n8n, and webhooks, some of the hardest concepts in the course.

Not everyone loved every part of it. One student mentioned the recorded video quality could have been better, fair feedback, and something I'm already fixing for batch two.

I'm two weeks into batch two now, the last one I'm running with ICT Bangla. Right after it wraps, I'll share the full close, what the whole run built in me, and why I'm stepping back after this.