Generative AI can make new tasks possible and old tasks better—but the first question is always whether you need it at all. Sometimes a simpler method is faster, cheaper, and more reliable.

When you do need it, complex applications are built from design patterns called archetypes. In this video, you'll learn how retrieval-augmented generation (RAG) grounds a model in your private knowledge by retrieving relevant context before it answers, and how AI agents go further, reasoning through a plan, using tools, accessing memory, and taking action in a loop until the goal is met. These patterns are the building blocks of the next generation of AI tools.