Generative AI has evolved from basic chatbots into agentic systems capable of much more than just conversation. This talk explores what it takes to build a production ready GenAI agent, which doesn't just answer questions, but understands context, uses tools, and takes actions to solve real problems. Even the smartest LLM is limited without access to up-to-date knowledge or the ability to interact with the world. We'll look at how augmenting a model with memory, external data, and action capabilities turns a passive chatbot into a proactive, helpful assistant.

We'll dive into what the modern building blocks are for building AI-powered software. Techniques like Retrieval-Augmented Generation (RAG), embeddings, fine-tuning, and targeted tool use help ground your AI in real-world context, improving both precision and reliability. You’ll learn about open standards like the Model Context Protocol (MCP), a standardized interface for external tools and data. Along the way, we’ll consider how to balance speed, accuracy, and cost when designing systems for real-world performance.

We’ll touch on other important concepts to be aware of to enable you to build the AI solutions you need. Prompt and context engineering are core techniques for correctly guiding model behavior and reducing hallucinations by providing just the right information at the right time. You'll leave with tools and inspiration for building GenAI solutions which are more than a wrapper on top of a mainstream LLM.