What Are AI Agents? Moving Beyond Chat

AI chat interfaces are reactive: you ask a question, the AI answers. AI agents are proactive: you give them a goal, and they plan and execute a sequence of actions to achieve it.

This is not a minor evolution. It represents a qualitatively different relationship between humans and AI systems.

The Difference Between Chat and Agents

• Chat AI:
• You provide input
• AI produces output
• You decide what to do with the output
• You initiate every interaction

• AI Agent:
• You provide a goal
• Agent plans a sequence of steps
• Agent takes actions (searches web, writes files, calls APIs, runs code)
• Agent evaluates results and adjusts
• Agent reports back when done (or asks for guidance when stuck)

The agent operates with a degree of autonomy that chat AI does not.

The Components of an AI Agent

Most AI agent architectures include:

A language model (the brain) The model that does the reasoning: planning steps, interpreting results, deciding what to do next.

Tools (the hands) Functions the agent can call: web search, file operations, database queries, API calls, code execution, email sending. The range of tools determines what the agent can actually do.

Memory How the agent retains context. Short-term memory (the current conversation context) is standard. Long-term memory (persistent storage across sessions) is more complex and less standardised.

Planning mechanism How the agent breaks a goal into steps. The simplest approach: ask the LLM to plan; execute each step; evaluate; continue. More sophisticated approaches use dedicated planning algorithms.

Observation loop After each action, the agent observes the result and decides on the next action. This loop continues until the goal is achieved or the agent determines it cannot proceed.

Why Agents Matter Now

Agents represent the transition from AI as a productivity tool to AI as a capable worker.

• Current capable agents can:
• Research a topic across multiple web sources and produce a synthesised report
• Write, test, and debug code with minimal human intervention
• Manage and respond to email within defined parameters
• Monitor conditions and trigger actions when thresholds are met
• Navigate websites and perform multi-step online tasks

Agents are moving from demos to production deployment. Understanding them is becoming important for anyone working in AI-adjacent roles.

What This Module Covers

• This module provides a conceptual and practical foundation for AI agents:
• How agent architectures work
• The major agent frameworks (AutoGPT, CrewAI, LangChain)
• Practical agent patterns you can build today
• The significant safety and reliability challenges agents introduce
• Where agents are genuinely useful vs. where they are overhyped