Generative AI on Microsoft Azure

Chapter 6. Agentic Systems

In this chapter, we will define what an agent is, when to use agents, and how you can build secure agentic systems for enterprise environments. We will outline the key Azure services that support agentic AI deployments, ensuring scalability, security, and efficiency in real-world applications.

At its foundation, an AI agent is an intelligent software system capable of autonomously pursuing goals, using its reasoning and planning capabilities to perform a series of predefined or adaptive actions within an environment. Unlike traditional programs that follow fixed instructions, an agent exhibits a degree of autonomy: it can operate and make decisions independently, without constant human oversight. Its defining strength lies in its reasoning and planning abilities, giving it the capacity to analyze multimodal information, interpret complex contexts, and generate coherent plans and strategies toward achieving its objectives. In doing so, the agent can develop and execute a structured sequence of actions to reach a goal, informed by memory or context from past interactions. Together, these elements transform the agent from a reactive system into a proactive, goal-directed entity capable of adapting to new situations and optimizing its decisions over time.

In practical terms, a generative AI (GenAI) agent integrates several core components that collectively enable these intelligent behaviors. At its center lies a foundation model, typically a large language ...