AI Monday Prague — What I Learned This Time

Another great event full of insights, ideas, and practical takeaways

This week I attended AI Monday in Prague, and just like the one in Jihlava, it was an excellent experience. Great atmosphere, inspiring conversations, and talks packed with real-world knowledge.

Here are the key things I took away from the evening.

🔧 1. The Master Prompt Defines the Whole App

From Jindrich Faborsky, I learned how important the master prompt really is.

The quality of your master prompt often determines the quality of the entire application.

In practice, this means:

defining the app structure clearly
describing the behaviour in detail
setting consistent rules from the start

It saves hours of work and leads to far better output from tools like Lovable, Macaly, or Cursor.

🧠 2. How to Build an AI Chatbot on Company Documentation

The team from EMARK shared practical insights into building chatbots on top of internal company documents.

Their talk covered:

preparing and structuring documents
choosing a vector database
handling retrieval and context
optimising the system for real deployment

Very useful for anyone integrating AI into customer support or internal knowledge systems.

🛠️ 3. Open-Source AI Tools and How Livesport Uses Them

The host company Livesport, with speaker Marek Keřka, introduced several open-source tools for AI — including LiteLLM (YC W23).

What stood out to me was how Livesport:

combines different models
uses n8n for automation
builds internal AI workflows at scale

A great behind-the-scenes look at AI in production inside a large, fast-moving company.

🙏 Final Thoughts

Another fantastic AI Monday, full of great speakers and valuable insights.

Big thanks to everyone involved — I left with new ideas, inspiration, and motivation to push my own projects even further.

I’m already looking forward to the next one!