Agentic AI - From Theory to Practice

Target Audience

Practical users and managers who want to fully harness the potential of agentic AI in their individual work and within their organizations.

Learning Goals

• What is AI, how do language models and agentic systems work, and what limitations do they have?

Topics

Introduction

We start the workshop with a brief welcome, an agenda overview, and a look at the current AI landscape.

• What types of AI systems currently exist?
• How do chatbots, coding assistants, document assistants, spreadsheet assistants differ — and what does this mean for everyday professional work?

The goal is to collectively refresh our overview of the AI landscape, get to know each other, and integrate participants' expectations into the workshop.

Keywords: AI landscape, chatbots, coding assistants, document assistants, spreadsheet assistants, coworkers, integrations

AI Technology Under the Hood

In this first session, we provide an intuitive foundational understanding of what is likely the greatest technological revolution in the history of humanity.

• How do artificial neural networks work?
• What are vision models and how do they work?
• How are (custom) AI models trained?
• Why is data quality important?
• How do language models work as neural networks?

We open the "black box" and transform mysteries into comprehensible relationships. You don't need to be a mechanic to drive a car, but a basic understanding of the technology is usually advantageous.

Keywords: deep learning and neural networks, models, weights, tokens, training, inference, vision, language

From Chatbots to Agents - Agentic Architecture in Detail

Deep learning as modern technology opens new possibilities. The great hype around AI has arrived with the sub-category of agentic systems: They are able to interact with their environment, which causes the possibilities to explode.

• What makes an LLM an agentic system?
• What types of agentic systems exist or are possible?
• What are the limitations of agentic systems?

In this session, we aim to build an understanding of the prerequisites and functioning of agentic systems, starting from the technology of large language models.

Keywords: LLM, agentic system, scaffolding, harness, context management, tools, MCP, skills, prompt

Interactive Discussion

This flexible session is guided by the group's needs and deepens the topics most relevant to the workshop.

• Which use cases are most relevant for your company?
• When does local vs. cloud AI make sense?
• Open vs. closed models — what fits our organization?
• How can we use AI responsibly?

The goal is to independently deepen the topics most relevant to your company, derive concrete insights for everyday work, and identify next steps.

Keywords: discussion, use cases, local AI vs cloud, responsible AI, next steps, brainstorming

Format

This workshop is typically conducted on site as a full day event.