From the course: Agentic AI: A Framework for Planning and Execution

What exactly is an agent?

From the course: Agentic AI: A Framework for Planning and Execution

Start my 1-month free trial Buy for my team

What exactly is an agent?

“

- In the rapidly evolving world of artificial intelligence, you've likely heard the term AI agent thrown around quite a bit, but what exactly is an agent and why should you care? Let's cut through the hype and get to the heart of what makes something an agent. At its core, an AI agent is a software system that can perceive its environment, make decisions, and take actions to achieve specific goals, all with varying degrees of autonomy. Unlike traditional software that follows rigid, predetermined paths where developers need to plan ahead for different contingencies and how to handle them, sometimes even contingencies that you can't possibly plan for, agents instead can adapt their behavior based on their observations and reflection on what they already know. That not only makes them more powerful, but also more flexible to meet your business needs. Typically, you design them to act on behalf of a user or an organization to accomplish tasks, to assist a human in doing their work. Let me give an example. Say you're a sales executive preparing for a sales call. To be effective, you need to do a lot of research ahead of time; reading up on the company, the person you're speaking with, past experiences others may have had with that company or person. To be effective, you need to do this well, but that's a lot of cognitive load that is not directly part of what you do. Each instance is different; different people, documents, interactions, website structures, all of that stuff. It's hard to write a rules-based program to figure all of this out for you. And here's where an agent can help. They can summarize documents and extract key insights. They can determine sentiment, what's our relationship like with this customer? Or they can even take vague instructions like, can you update the Johnson account and send them our latest metrics; breaking all of this down into a tangible workflow. But how do they do this? Well, some of the terms that you're going to hear are autonomy. Agents can operate independently, making decisions without constant human guidance. Goal orientation. Agents can work towards specific objectives, whether that's simple like scheduling a meeting or answering customer inquiries, or more complex like optimizing a supply chain. Reactivity. Agents can sense changes in their environment and respond accordingly. Persistence. Agents can maintain state over time, learning from past interactions to improve future performance. In essence, agents are AI systems that go beyond simply processing information to actually making decisions and taking actions in pursuit of specific goals. In the next video, we'll explore how agents differ from traditional AI models and we'll help you better understand where they fit into the technology ecosystem.

Contents