A too short introduction to agentic AI
Let’s start with the bad news. In 90 minutes we will not make you an expert in LLMs or agentic AI. Afterwards you will not be able to explain the mathematics behind the attention mechanism. And we cannot predict how AI will precisely impact your business.
But we can help you predict how AI can improve your business. In an intense, knowledge-oriented talk we demystify what agents are and how they work under the hood, and therefore what their strengths, their weaknesses and their security implications are. No marketing hype, no overselling of the benefits. A down to earth primer on the technology behind today’s AI revolution. Without using formulas, we promise.
What you will learn
- The fundamental workings of an LLM. What it does when it produces a word, and what it therefore cannot do.
- The difference between a chatbot and an agent, and what a harness, a skill and an MCP server are.
- The main security frameworks used to keep agents safe, from banning AI outright to sandboxing an agent inside your own application.
- The architectures used to deliver business value with agents, and where each one puts your data.
Why this one is different
It comes out of practice. We have run agentic development in our own production work since 2024, and we wrote our own methodology for keeping quality, security and architecture under control when code arrives faster than anyone can read it.
We cite the measurements. Where we make a productivity claim we name the study, including the ones that found negative effects and the surveys where most enterprise pilots show no impact on the P&L at all.
We explain the mechanism. Attention gets a five step walkthrough on a real sentence, with no formulas. Security is in the news every week, so the room already knows the words. What it has never been shown is why prompt injection cannot simply be patched out.
Practical
| Length | 90 minutes, a 60 minute version is available |
| Language | English or Dutch |
| For | Executives, managers and decision makers with some IT background |
| Prerequisites | None. Some experience with ChatGPT helps. No coding, no mathematics |
| Format | Talk with questions. No hands-on |
| Delivered by | Frank Dekervel or Rein Lemmens, solo or together |
The speakers
Frank Dekervel, partner and CTO. Runs agentic development at Kapernikov and wrote the methodology behind it.
Rein Lemmens, CEO and co-founder. Works with infrastructure managers on the gap between what a technology can do and what an organisation can adopt.
See for yourself
Over 100 people have been through our agentic AI trainings, in the developer and the executive edition.
Read the interview with Frank: Agentic AI, Applied
See the slides: the executive deck is public, and it carries the argument better than this page will.
Want the longer version instead? The two and a half hour training for knowledge workers covers the same ground with room for deployment strategy, the EU AI Act and the impact on HR.
Want this session for your event or your management team?
Tell us what you are planning and we will come back to you about the date. Prefer email? Write to frank@kapernikov.com.