> ## Documentation Index
> Fetch the complete documentation index at: https://academy.pathfindr.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# What Makes an AI Agent

> Understand the difference between chatbots, copilots, and agents, and when each one is the right choice.

export const Quiz = ({question, options, correctIndex, explanation}) => {
  const [selected, setSelected] = useState(null);
  const [showResult, setShowResult] = useState(false);
  const handleSelect = index => {
    if (showResult) return;
    setSelected(index);
    setShowResult(true);
  };
  const isCorrect = selected === correctIndex;
  return <div className="my-8 p-6 rounded-2xl border border-zinc-200 dark:border-zinc-700 bg-gradient-to-br from-zinc-50 to-white dark:from-zinc-900 dark:to-zinc-800 shadow-sm">
      <div className="flex items-center gap-3 mb-4">
        <div className="flex items-center justify-center w-10 h-10 rounded-xl bg-blue-100 dark:bg-blue-900/40">
          <span className="text-xl">📝</span>
        </div>
        <div className="flex flex-col">
          <p className="font-semibold text-zinc-800 dark:text-zinc-100 m-0">Knowledge Check</p>
          <p className="text-xs text-zinc-500 dark:text-zinc-400 mt-0.5 m-0">Select the best answer</p>
      </div>
      </div>
      
      <p className="text-zinc-700 dark:text-zinc-200 font-medium mb-5 text-lg leading-relaxed">{question}</p>
      
      <div className="space-y-3">
        {options.map((option, index) => {
    const isSelected = selected === index;
    const isCorrectOption = index === correctIndex;
    let optionStyles = "border-zinc-200 dark:border-zinc-600 hover:border-blue-400 hover:bg-blue-50 dark:hover:bg-blue-900/20";
    if (showResult) {
      if (isCorrectOption) {
        optionStyles = "border-green-500 bg-green-50 dark:bg-green-900/30";
      } else if (isSelected && !isCorrectOption) {
        optionStyles = "border-red-500 bg-red-50 dark:bg-red-900/30";
      } else {
        optionStyles = "border-zinc-200 dark:border-zinc-700 opacity-60";
      }
    }
    return <button key={index} onClick={() => handleSelect(index)} disabled={showResult} className={`w-full text-left p-4 rounded-xl border-2 transition-all duration-200 ${optionStyles} ${!showResult && 'cursor-pointer'} ${showResult && 'cursor-default'}`}>
              <div className="flex items-start gap-3">
                <div className={`flex-shrink-0 w-6 h-6 rounded-full border-2 flex items-center justify-center text-xs font-bold ${showResult && isCorrectOption ? 'border-green-500 bg-green-500 text-white' : showResult && isSelected && !isCorrectOption ? 'border-red-500 bg-red-500 text-white' : 'border-zinc-300 dark:border-zinc-500 text-zinc-500 dark:text-zinc-400'}`}>
                  {showResult && isCorrectOption ? '✓' : showResult && isSelected ? '✗' : String.fromCharCode(65 + index)}
                </div>
                <span className="text-sm text-zinc-700 dark:text-zinc-200 leading-relaxed">{option}</span>
              </div>
            </button>;
  })}
      </div>
      
      {showResult && <div className={`mt-5 p-4 rounded-xl ${isCorrect ? 'bg-green-100 dark:bg-green-900/40 border border-green-200 dark:border-green-800' : 'bg-amber-100 dark:bg-amber-900/40 border border-amber-200 dark:border-amber-800'}`}>
          <div className="flex items-start gap-3">
            <span className="text-xl">{isCorrect ? '🎉' : '💡'}</span>
            <div className="flex flex-col">
              <p className={`text-sm font-semibold m-0 mb-1 ${isCorrect ? 'text-green-800 dark:text-green-200' : 'text-amber-800 dark:text-amber-200'}`}>
                {isCorrect ? 'Correct!' : 'Not quite right'}
              </p>
              <p className="text-sm text-zinc-600 dark:text-zinc-300 m-0 leading-relaxed">{explanation}</p>
          </div>
          </div>
        </div>}
    </div>;
};

<img src="https://mintcdn.com/pathfindr/vxED2dxssEHZg3yn/images/understandingAgents.webp?fit=max&auto=format&n=vxED2dxssEHZg3yn&q=85&s=68d28e4e0dd826bfd14d1524d83d6ec5" alt="understanding AI agents" noZoom={true} width="1280" height="720" data-path="images/understandingAgents.webp" />

## What you will learn

Most people using AI today are still in chatbot mode. They open a tool, type a question, get an answer, and close the tab. That works, but it caps how much value you can extract.

This section shows you the three levels of working with AI and what it actually takes to move into agent territory.

<CardGroup cols={4}>
  <Card title="Three levels of AI" icon="square-1">
    Distinguish between chatbot, copilot, and agent modes of working
  </Card>

  <Card title="What qualifies as an agent" icon="square-2">
    Identify the three qualities that separate agents from simpler AI setups
  </Card>

  <Card title="Agent capabilities" icon="square-3">
    See how the Thinker, Assistant, and Creator roles apply when agents run autonomously
  </Card>

  <Card title="Right tool, right job" icon="square-4">
    Know when an agent is the right solution and when a simpler approach is better
  </Card>
</CardGroup>

***

## The three levels of working with AI

Think of these as three different ways to work with a colleague, each with a different level of independence.

### Level 1: Chatbot

You control every interaction. You ask a question. You get a response. You ask another question.

The AI has no memory of your preferences, no access to your tools, and no ability to act without you prompting it. This is where most people start, and it is genuinely useful for brainstorming, quick research, and one-off tasks.

<Info>
  But you are the engine. Nothing happens without you in the chair.
</Info>

### Level 2: Copilot

You are still in the driving seat, but the AI knows more about you. You have given it context through a custom GPT, a Project in Claude, or a set of uploaded documents. It has preferences, tone guidelines, and background knowledge baked in.

This is where a setup like a contract review assistant or a communications helper sits. You have built something reusable. But you still trigger every action.

<Tip title="Where are you right now?">
  If you have set up custom GPTs, Projects, or Gems in Weeks 1 and 2, you are operating at Level 2. That is a strong foundation. Agents build directly on that same skill (writing clear instructions, providing context, specifying outputs) but add automation on top.
</Tip>

### Level 3: Agent

The AI works without you prompting it. It triggers itself based on conditions you set. It connects to your tools: email, calendar, documents, web data.

It does this repeatedly, following the same process each time, improving as you give it feedback. You shift from doing the work to reviewing the work. From being the operator to being the manager.

<CardGroup cols={2}>
  <Card title="Copilot" icon="car-side">
    You drive. AI assists. Every action requires your input.
  </Card>

  <Card title="Agent" icon="robot">
    You set the rules. AI drives. You review the results.
  </Card>
</CardGroup>

<Accordion title="Test your understanding">
  <Quiz
    question="You have built a custom GPT in ChatGPT that reviews contracts against your company's standard terms. You upload a contract, it produces a departure table. What level is this?"
    options={[
  "Chatbot: you are just asking it questions",
  "Copilot: you have built in context and preferences, but you trigger every action",
  "Agent: it is doing the work automatically",
  "Agent: it has access to your documents"
]}
    correctIndex={1}
    explanation="This is a copilot. You have given it a clear role, uploaded reference documents, and built a reusable setup. But you still trigger every interaction manually. To make this an agent, you would need it to automatically detect when a new contract arrives and process it without your prompt."
  />
</Accordion>

***

## The three qualities of an agent

Not everything that calls itself an agent actually is one. The market is noisy, and plenty of tools use the word "agent" as a label for what is really a chatbot with extra steps. Here is how to tell the difference.

<CardGroup cols={4}>
  <Card title="Proactive" icon="bolt">
    It initiates work on its own based on triggers you define. You do not have to prompt every time. A schedule fires, a file lands in a folder, an email arrives, and the agent starts working.
  </Card>

  <Card title="Integrated" icon="plug">
    It connects to your actual tools and data. Email, calendar, documents, CRM, project management. The more integrated it is, the more it can do on your behalf.
  </Card>

  <Card title="Repeatable" icon="arrows-rotate">
    It follows the same process each time, producing consistent and predictable outputs. You can give it feedback and improve its performance over time.
  </Card>

  <Card title="Verifiable" icon="eye">
    You can see what it is going to do before it does it. Flowcharts, logs, and human-in-the-loop review give you visibility and control over every action.
  </Card>
</CardGroup>

If something you are using requires you to prompt it every time, it is not proactive. If it cannot connect to your tools, it is not integrated. If it gives you different quality results each time with no way to refine it, it is not repeatable.

<Tip>
  Those are copilots, and they are valuable. But they are not agents.
</Tip>

***

## What agents can do

Agents draw on the same three roles you have already practised (Thinker, Assistant, and Creator) but they apply those roles autonomously.

<AccordionGroup>
  <Accordion title="Agent as Thinker" icon="lightbulb">
    An agent in Thinker mode researches, analyses, and surfaces insights without you asking. A competitive analysis agent scans competitor LinkedIn profiles every Monday and sends you a briefing. A market research agent tracks industry publications and flags relevant developments. The thinking happens on schedule, not on demand.
  </Accordion>

  <Accordion title="Agent as Assistant" icon="list-check">
    An agent in Assistant mode handles recurring operational tasks. It triages your inbox, sends meeting briefings 24 hours before each calendar event, follows up with attendees who have not responded, or files completed documents into the correct folder in your Drive. The admin work runs in the background.
  </Accordion>

  <Accordion title="Agent as Creator" icon="pen-nib">
    An agent in Creator mode generates content on a schedule or in response to triggers. It drafts weekly reports from your data, writes social media posts based on company updates, or creates first-draft blog content from research it has already gathered. The creative output arrives ready for your review.
  </Accordion>
</AccordionGroup>

***

## When an agent is not the right answer

Not every task needs an agent. If you are doing something once, a chatbot is fine. If you need flexibility and judgement each time, a copilot with strong context is often better.

Agents work best when the task is repetitive, the process is definable, and the triggers are clear. If you cannot describe the steps, the trigger, and the expected output, you are not ready to build an agent for that task yet.

<Warning title="A common mistake">
  Jumping straight into building an agent before you understand the work it needs to do. The most valuable step is often the least exciting: mapping out what you currently do, step by step, before you automate any of it.
</Warning>

<Accordion title="Test your understanding">
  <Quiz
    question="Your team wants an AI that monitors competitor social media every week and sends a summary to Slack. Which three qualities must this have to qualify as an agent?"
    options={[
  "Fast, accurate, and cheap",
  "Proactive, integrated, and repeatable",
  "Creative, analytical, and collaborative",
  "Automated, intelligent, and scalable"
]}
    correctIndex={1}
    explanation="An agent must be proactive (triggers on schedule without you prompting), integrated (connects to LinkedIn, Slack, and any other tools it needs), and repeatable (follows the same process each week with consistent quality you can refine over time)."
  />
</Accordion>

***

## Quick checkpoint

You are done with this section when you can:

<CardGroup cols={4}>
  <Card title="Name the three levels" icon="circle-check">
    Explain the difference between chatbot, copilot, and agent to a colleague
  </Card>

  <Card title="Spot the qualities" icon="circle-check">
    Identify whether a tool is proactive, integrated, and repeatable, or not
  </Card>

  <Card title="Match the role" icon="circle-check">
    Describe how Thinker, Assistant, and Creator roles work in an agent context
  </Card>

  <Card title="Choose wisely" icon="circle-check">
    Decide whether a task is better suited to a chatbot, copilot, or agent
  </Card>
</CardGroup>

<div className="mt-8" />

<Card title="Next: Designing Your AI Agent" icon="arrow-right" href="/leadership-cohort/week-4-build-your-own-ai-agents/designing-your-agent">
  Learn how to map your workflows, identify where agents fit, and write an agent job description
</Card>