This talk offers an in-depth exploration of the evolution of language model usage—starting from traditional large language models (LLMs) to the practical design and deployment of agentic AI systems. We’ll clarify the distinction between Agentic AI and AI agents, and demonstrate how agentic patterns enhance reasoning, tool use, and autonomy. Using Google ADK, I’ll showcase how to build and run single-agent setups, as well as more advanced workflows involving parallel and sequential agents. This hands-on walkthrough will illustrate how agents can be composed to solve increasingly complex tasks efficiently and reliably.