Table of Contents

1) AI Agents

You’ve chatted with ChatGPT, it answers your question, and then stops. But what if the AI didn’t stop? What if it could go out, open a spreadsheet, run a search, write a report, and send you the summary.

That’s an AI Agent, an AI that not only thinks, but also acts.”

1. Perception — The agent reads the world: user inputs, API responses, file contents, or sensor data.
2. Reasoning — The LLM plans: “To achieve this goal, I’ll need to do steps A → B → C.”
3. Action — The agent executes: calling APIs, running code, or writing to external systems.
4. Reflection / Memory — It evaluates results, stores what it learned, and loops until the goal is done.

AI agents are systems that decide what to do next on their own. They perceive inputs (text, files, tools, environment state), plan multi-step actions, call tools/APIs, and update memory until the goal is met.

Agentic AI

This term typically refers to a more advanced form of AI that embodies more autonomy and decision-making capabilities.

Examples