Build your first production-ready agentic workflow
Practice day for IT managers, architects and consultants who want to build an agentic workflow — with LangGraph or n8n, with real error handling and human escalation path. You go home with working code.
What you'll learn
-
State machine vs. orchestrator vs. reactive
When to choose which pattern — and how that choice affects scalability and debuggability.
-
Tools, API integrations and state management
How to make external tools (REST, databases, email) available to agents without security risks.
-
Error handling and retry strategies
Production-ready error handling: exponential backoff, dead-letter queues and graceful degradation.
-
Building a human escalation path
How to build a human-in-the-loop that catches the right moment based on confidence, process context or time pressure.
Programme
A day in detail — no surprises.
| # | Time | Module | Content |
|---|---|---|---|
| — | 9:00–9:30 | Setup and architecture choice | Tool choice, environment setup, pattern selection |
| — | 9:30–11:30 | Build core agent | State graph, tool integrations, base flow |
| — | 11:30–12:30 | Error handling and observability | Retry logic, logging and monitoring |
| — | 13:30–15:00 | Build human escalation | Approval flow, notification and audit trail |
| — | 15:00–16:00 | Code review and production readiness | Peer review, security checklist, next steps |
Immediate results
- Working agentic workflow ready to deploy
- Reproducible architecture patterns for the team
Long-term
- Build internal implementation capacity
- Shorter time-to-production for agentic BPM projects
Who it's for
- IT managers and digital leads
- Enterprise and solution architects
- Technically experienced consultants
About the trainer
Wiemer Kuik
Founder & senior agentic BPM advisor
Wiemer Kuik has worked at the intersection of process management, organisational change and technology for 40+ years. As a former transformation advisor at major Dutch organisations, he knows what it costs when processes don't work — and what it delivers when they do. His focus is on the practical side of agentic BPM: not the hype story, but the architectural choices, governance setup and adoption approach that determine whether an implementation actually succeeds.
LinkedInParticipant experiences
After one day I had a working agent in our test environment. That is the best proof that the approach works — hands-on, no slides.
The error handling and retry logic covered here are exactly the things you don't find in the official documentation. Pure practical knowledge.
I have been building on agent frameworks for years, but the human escalation path was always a blind spot for me. Now I know how to set it up properly.
The peer review at the end was gold. Colleagues who build along on the same day see things you don't. Such a learning setting really works.
I rarely start implementing the next day based on what I learned in a training. With this training that was literally the case.
Not sure about the fit?
Book an intake — we'll spend 30 minutes on your processes and give you honest advice.