Different types of AI agents
AI agents come in varying levels of complexity and capability, ranging from simple retrieval-based agents to fully autonomous systems. Understanding these different types helps organizations and developers select the right kind of AI agent for specific use cases. In this section, we are going to cluster AI agents into three primary types: retrieval agents, task agents, and autonomous agents.
Retrieval agents
In Chapter 1, we introduced the concept of RAG as a technique in GenAI applications where an LLM retrieves relevant documents or snippets from a knowledge base (properly embedded and stored in a VectorDB) before generating responses.Retrieval AI agents build upon the foundations of RAG but incorporate advanced agentic behaviors, making them more autonomous and adaptive. In fact, we are adding to a standard RAG pipeline an additional layer of intelligence and planning that allows the agent to “strategize” on how to retrieve the most relevant...
