YUKEN
WorkIndustry Rockstar

Build along bootcamps where every learner leaves with a working agent

Industry Rockstar asked us to design and deliver two live build along bootcamps: a Make AI Agents programme and a Building AI Agents in n8n programme. Every demo agent was built and deployed to a live tenant, so learners left with working automations rather than screenshots.

Client

Industry Rockstar

Year

2026

Engines

Solutions

Engagement

Two bootcamps, designed and delivered

Impact

Around 20 hours of live build along delivery, with every demo agent deployed to a live tenant

Industry Rockstar runs live business and technology programmes for entrepreneur audiences worldwide.

The situation

Build along training only works if the demonstrations work live in front of an audience. Two bootcamps, around 20 hours of combined delivery, needed designing, scripting and running by someone who could build every agent live and recover from problems mid session.

What we did

Build the system

01

Designed both programmes in full

Session structure, decks, facilitator guides, student guides and delivery scripts running to around 8,000 words per session.

02

Built every demo before the room saw it

Every demo agent was built and deployed to a live cloud tenant ahead of delivery: multi tool agents with human in the loop approval, structured JSON output and vector store retrieval, all running on real infrastructure.

03

Engineered the delivery environment

We wrote custom Python against the n8n API to provision and test workflows at scale, so each cohort started from a known good environment.

04

Delivered the build alongs live

Learners built alongside us session by session and left with working automations they had made themselves.

Outcomes

What the client was left with

  • Two full bootcamp programmes designed and facilitated
  • Multi tool agents with human in the loop and structured output
  • Custom Python against the n8n API to provision and test environments
  • Delivery scripts running to around 8,000 words per session
Built with

Make, n8n, OpenAI, Python, Vector stores

Have something similar in mind?

Describe the situation and we will scope the work.