Only 12 seats
In 4 weeks, you'll have a real AI team running in your work or business. Hands-on. No coding. Sessions held in person in Hammarby Sjöstad, Stockholm.
Bootcamp is held in English
From the Workshop Floor
YOUR FIRST
AGENTIC AI
SYSTEM
Workshop #5 · 2025
No theory. We build a real-world AI project.
What You Will Build
In 4 weeks, you deploy a multi-agent AI system, one orchestrator and four agents. Every action coordinated, traceable, and defined by you.
00
Orchestrator Agent
01
Reply Agent
02
Follow-Up Agent
03
Outreach Agent
04
Performance Agent
Tech Stack
Every tool was chosen for a reason. The right one for non-technical founders.
Claude Sonnet 4.6
Why not GPT-4? Fewer hallucinated decisions. Handles long system prompts reliably critical when agents send emails.
n8n
Why n8n? n8n runs on infrastructure you control, no task limits, no surprise pricing, no outages you didn't cause. More reliable, fully yours.
Airtable
Why Airtable? Every AI decision is logged and readable. See every state change without touching code.
Claude Code
Why it matters: When your agent behaves unexpectedly, describe the problem in plain English. Claude Code diagnoses and fixes it, no developer needed.
Live Scenario
HOW IT WORKS
Incoming email · Lead: Erik Lindström
"We're interested, but the price feels a bit high for us right now. Also, we're not sure whether we can move forward with implementation this month."Signals detected by Claude AI
ORCHESTRATOR DECISION
Standard lead
Draft a measured reply. Acknowledge price concern, don't push. Move conversation forward.
Agent Output · Inbound Reply Agent
"Hi Erik! Totally understand on timing, no rush this month. On pricing, I'm happy to walk through options. Want to jump on a quick call?"High-value lead
Route to human review. Message is sensitive and worth personal attention.
→ Human reviewIf they go quiet
System tracks state. Suggests follow-up at the right time, no blind nudging.
→ State trackedWho It's For
This is for you if this sounds familiar:
You don't need to
code.
You need to think
in systems.
How It Works
Each week produces a working component. No passive learningyou ship something every session. One in-person build session per week (3 hours). 12 hours in-person total across 4 weeks, plus 1 hour of individual practice per week.
Week 1
System Design & Decision Logic
Before you build anything, you define the rules. What the AI sends. What it escalates. What it ignores. No guessing later.
Week 2
Build the Inbound Reply Agent
Your inbox goes live. Real emails come in. The agent reads, classifies, and replies in your tone, within your rules.
Week 3
Follow-Up Agent + AI Quality Auditor
Missed follow-ups kill deals. This agent tracks every thread, knows when to nudge, and documents every decision it makes.
Week 4
Test, Harden, Deploy
You break it before the real world does. Fix the edge cases. Then deploy to your actual inbox and run it live.
Images from our workshops
Agentic AI Bootcamp · Stockholm · Workshops 01–05
← Swipe to see more →
What people say
Average rating
Came in skeptical, left with a working agent in my inbox. By hour two I'd built the send loop myself. Not following along. Actually doing it.
When the agent read an email and sent a draft back for my review, it finally clicked. I'd been trying to understand ReAct for months.
Structured and hands-on from start to finish. I left with a working AI agent, no theory, no waiting. First time it actually felt real in my business.
Your instructor
Javid Valiyev
Founder · Genaition
Javid builds Applied AI and teaches people and businesses how to implement them without needing a developer. His workshops are known for being hands-on, structured, and beginner-safe. You don't leave without a working system.
"You don't leave without a working system. That's the rule."
Apply
4,985 SEK
3,985 SEK
Save 1,000 SEK until 5 July
No commitment. We'll reach out within 24 hours.
FAQ
English
Svenska
