Putting Business Knowledge to Work
How do you turn internal documentation, Confluence pages, PDFs, and even codebases into a useful knowledge assistant? In this TechTalk, we’ll explore how Retrieval-Augmented Generation (RAG) makes this possible. You’ll learn how RAG works in practice, see realistic use cases, get architecture tips, and find out which pitfalls to avoid.
Retrieval-Augmented Generation
Generative AI is changing the way we work. Large Language Models are capable of a great deal, but they also have clear limitations. They can hallucinate and only know what they learned during training, which means they lack up-to-date information. RAG offers the solution. It connects models to your own knowledge, so they stop guessing and deliver reliable results.
Walter Heck, Technology Executive and Co-Founder and CTO of Helixiora, takes you through the concepts, challenges, and opportunities of RAG. From vector databases and embedding algorithms to GraphRAG and tool calling, he shows you how to bridge the gap between an abundance of information and actionable insights.
This 45-minute session combines clear explanations with practical examples and gives you concrete ideas that you can immediately apply to your own projects.
Agenda
- A brief introduction to Large Language Models
- The limitations of LLMs: hallucination and knowledge cutoff
- Retrieval-Augmented Generation as a response to these limitations
- How RAG works in practice
- Real-life examples of RAG
- Vector databases and vector search: chunking and embedding algorithms
- GraphRAG and its capabilities
- The limitations of RAG
- Other concepts on the horizon: tool calling and MCP
Meet our speaker
Walter Heck
Technology Executive and Co-Founder and CTO at Helixiora
Walter has more than 20 years of experience helping organizations adopt a data-driven approach using cloud technologies. He has guided organizations through complex digital transformations, developed data-driven strategies, and embraced new technologies to drive growth and innovation. Today, as a co-founder of Helixiora, he helps organizations unlock the true potential of their data.
