What Makes an Agent Agentic?
1. What Makes an Agent Agentic?
Welcome back! In the last lesson, we learned that AI agents combine reasoning, planning, and action to achieve goals. But what exactly gives Agents these capabilities? What makes an agent... agentic? Let's peek under the hood and discover the core components of AI agents.2. Jeeves: Our AI Travel Agent
Previously, we were introduced to Jeeves. An AI travel agent that, based on our request, looked at our itinerary, booked appropriate flights and hotels, and presented us with a detailed travel plan. Let's trace through what happens when you prompt a system like Jeeves.3. The Agentic Trinity: Model, Tools, and Orchestration
When you typed your prompt to book travel plans, three essential components worked together: The model, tools, and the orchestration layer.4. The Agentic Trinity: Model
First, there's the Model. Think of this as Jeeves' brain. This is a large language model that understood your request and figured out what steps to take.5. The Agentic Trinity: Model
It reasons through the presented problem and breaks it down into smaller steps. In this example, it takes a travel request,6. The Agentic Trinity: Model
and breaks it down in the different steps the agent needs to take. Without this, Jeeves would not be able to develop the necessary steps to help him achieve his objective.7. The Agentic Trinity: Tools
That said, even the smartest brain needs a way to interact with the world. That's where Tools come in.8. The Agentic Trinity: Tools
To help build your travel plans, Jeeves accessed your calendar,9. The Agentic Trinity: Tools
checked company travel policies,10. The Agentic Trinity: Tools
and searched Expedia and Booking.com for the best flights and hotels. Each of these is a tool. Tools extend the model’s ability to get up-to-date information, perform actions, and interact with digital interfaces. Tools are essential for bridging the gap between the model and the real world.11. The Agentic Trinity: Orchestration
Next up, we have the orchestration layer. The orchestration layer is a continuous loop that controls how an agent processes information, remembers information, and makes decisions. Think of it as managing the agent's decision-making cycle: it takes in data, thinks about what it means, and decides what to do next. It also keeps track of everything the agent has done so far. This cycle spins until the agent achieves its goal or hits a predetermined stopping point.12. The Agentic Trinity: Orchestration
This orchestration layer can be simple, like following basic if-then rules, or more sophisticated, involving complex reasoning chains and even other AI models. We'll spend more time on orchestration in the next chapter. For this lesson, what you need to know is that it keeps the agent going until the goal is achieved or a stopping point is reached.13. Let's Practice!
We’ll focus on these components in more depth, but now — let’s test your skills.Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.