Seven workshop-style lessons, about 6 hours: an AI certification course for a Swiss online institute
A Swiss online institute runs an AI certification for students and young professionals, built to make them useful with AI at work. We wrote and recorded seven workshop-style lessons for four of its modules, about 6 hours of video, taking students from custom GPTs and workspace AI to automations, AI agents and vibe-coded apps.
- Custom LLMs
- Content
- Automation
- Apps
Snapshot
- Client type
- Online institute with degree and certification programs
- Country
- Switzerland
- Industry
- Education
- System
- AI Training (recorded course)
- Use cases
- AI training and upskilling
- Stack
- Tools taught include Gemini, ChatGPT, Claude, Perplexity Comet, Leonardo AI, Gamma, HeyGen, Make.com, Voiceflow, Airtable, Bolt, Lovable, Supabase
- Timeline
- 8–30 November 2025
- Headline numbers
- 7 lessons · 5 hr 58 min of video · 4 program modules · work for hire
What did the client need?
Practical lessons, recorded by someone who builds with these tools every day, that students can follow and repeat.What did we build?
Seven screen-recorded, workshop-style lessons mapped to four modules, each with written guidelines or a guide.How does it run day to day?
Self-paced and fully online, inside the institute's certification program.
What did the client need?
The certification is meant to make students faster in their daily work, able to create professional content, automate processes and build simple AI tools of their own. The program outline set what each module had to cover and a minimum video length.
AI tools change month to month, so slides about features date quickly. Students need to see real builds, including what goes wrong, to repeat them on their own. The institute needed a practitioner to write and present the lessons, with resources students could follow along with.
What did we build?
| Program module | Lessons |
|---|---|
| Productivity in daily work | 1. Customising LLMs · 2. Workspace AI |
| Professional content and brand building | 3. AI image generation and multimedia design · 4. AI video generation |
| Automation: workflows, agents and APIs | 5. APIs and webhooks · 6. Workshop: no-code automations and AI agents |
| AI for building a business | 7. Vibe coding |
Most lessons are live builds: we set up the tools, build something that works and fix problems on camera, with only as much theory as each build needs. Each lesson came with written guidelines, and some with supporting resources students reuse across several lessons.
How does it run day to day?
The lessons run inside the institute's self-paced online program. Students work through them in order, and each lesson builds on the one before: material created in one lesson becomes the starting point for a build in a later one. The course was made as work for hire, so the lessons and resources belong to the institute.
What were the results?
We delivered seven recorded lessons, 5 hours 58 minutes of video in total, covering four of the program's modules, written and recorded between 8 and 30 November 2025. We don't have student outcome figures to publish; the course belongs to the institute.
- Recorded lessonsDelivered November 2025
- 5 hr 58 minOf video in totalDelivered November 2025
- Program modules coveredDelivered November 2025
What did we learn, and what are the limits?
- Teach with demo businesses, not real client data. Every build uses a made-up company or public information, so no client data ends up in a recording.
- Build live and keep the mistakes in. Errors, outages and wrong settings were fixed on camera, so students see how to recover on their own builds.
- Reuse one set of materials across lessons. Each lesson builds on the last, which makes the course feel like one project rather than seven demos.
- Explain technical ideas with everyday analogies, so students without a technical background can follow.