"Alright everyone, look at the title on the screen — Agentic AI: Goal-Driven Intelligence That Acts.

Now, before we dive in, I want you to notice one important word here — acts. Traditional AI responds. Agentic AI acts. That single word changes everything about how these systems work.

In the next few minutes, we'll break this down into six clear parts using this infographic. By the end, you'll understand not just what Agentic AI is, but also when to use it and why it matters."

"Let's start with the basics. Look at section one on the top left.

Agentic AI receives a goal — not step-by-step instructions.

Think about the difference. When you use ChatGPT to write an email, you're giving it a task. But when you use an Agentic AI, you're giving it a mission. You say: 'Here's what I want to achieve,' and the AI figures out the how by itself.

Notice the Core Idea highlighted in the green box — 'Instead of waiting for commands, an agent pursues a goal, plans, acts, learns, and adapts like a proactive teammate.'

That word teammate is important. An agent isn't a tool you operate. It's more like a colleague you delegate to."

"Now look at the top right. This is one of the most misunderstood ideas about Agentic AI.

Not every problem needs an agent.

If your task is simple — like writing an email or summarizing a paragraph — a normal prompt is enough.

If your task is fully rule-based — like moving invoices from one folder to another — traditional automation is better.

Agentic AI is for the in-between problems. Look at these five conditions on the screen:

1. Clear Goal — you know what outcome you want.
2. Multi-Step — it can't be solved in one step.
3. Uncertainty — the situation may change.
4. Multiple Tools — it needs different apps and systems.
5. Not Rule-Based — you can't write down every step in advance.

If your problem checks these boxes — as shown in the Ideal for box at the bottom — that's when Agentic AI shines."

"Now, look at this beautiful circular diagram in section three. Every AI agent has four core capabilities. Think of these as the four senses of an intelligent worker.

Perception — the agent observes the world. It checks weather, prices, live data, anything.

Reasoning — the agent thinks. It plans, evaluates options, and decides what to do.

Action — the agent acts. It uses tools, sends emails, books flights, fills forms.

Memory — the agent remembers. Short-term for the current task, long-term for your preferences.

Now here's the key insight — look at the highlighted box at the bottom. 'What makes it agentic? The combination of reasoning + action.'

Perception and memory already exist in normal software. But when a system can think and then do — that's where the magic happens."

"Now, look at the middle section. This is one of the most practical parts of understanding Agentic AI.

Autonomy is not a yes-or-no switch. It's a spectrum.

On the left — Low Autonomy — the AI only suggests. The human decides everything.

In the middle — Medium Autonomy — the AI plans and prepares, but waits for human approval.

On the right — High Autonomy — the AI acts independently to achieve the goal.

Notice the arrow at the bottom — More Human Control on one side, More AI Autonomy on the other.

Here's the practical wisdom: in most real-world business systems today, we operate somewhere in the middle. Why? Because humans want to review before big actions like spending money, sending customer emails, or making irreversible decisions."

"Now let's make this real with an example. Look at section five.

Imagine you tell an agent: 'Plan a 7-day vacation in coastal Italy for 4 people.' But it's not that simple — one person is vegetarian, two leave on day five, one hates boats, no chain hotels, and you love antique shops.

This is a messy, human goal. Traditional software would break instantly.

But look at how the agent works — follow the arrows:

• Research → It finds destinations, hotels, restaurants.
• Plan & Adapt → It builds an itinerary and adjusts to constraints.
• Act → It creates the plan, reserves options.
• Review & Approve → It presents everything to you for approval.

And notice the bottom box — Human-in-the-Loop. You decide how much freedom to give the agent. That's the beauty of Agentic AI — you stay in control."

"Now look at the comparison table at the bottom left. This is a summary you should remember for your exams and interviews.

• Input: Traditional software needs instructions. Agents need only a goal.
• Execution: Traditional software follows fixed steps. Agents plan and adapt.
• Adaptability: Traditional software is stuck in rules. Agents adjust to new information.
• Tool Use: Traditional software uses one tool. Agents use many.
• Nature: Traditional software is reactive. Agents are reactive and goal-driven.
• Outcome: Traditional software completes tasks. Agents achieve outcomes.

Read this line once again — achieves the desired outcome. That's the philosophy of Agentic AI."

"And finally, the Key Takeaway on the bottom right — read this with me:

Agentic AI is goal-directed intelligence that perceives, reasons, acts, and remembers. It pursues goals with a chosen level of autonomy, adapting along the way — turning complex, uncertain objectives into real-world outcomes.

And remember these four qualities shown at the bottom:

• Goal-Directed — focused on outcomes.
• Adaptive — changes plans as needed.
• Tool-Enabled — uses multiple systems.
• Collaborative — works with humans, not instead of them.

"So, students, next time someone asks you what Agentic AI really is, don't just say 'AI that does tasks.'

Say this instead:

'Traditional software follows instructions. Agentic AI pursues goals.'

That one sentence captures the entire revolution we are living through right now.

Any questions before we move to the next topic?"