Organizing the Cloud

When you log into the AWS Console, you are met with hundreds of services. For the AWS Certified AI Practitioner exam, you must be able to instantly categorize a service into one of three "Kingdoms."

If you confuse an Analytics service with an AI service, you will select the wrong architectural answer. In this lesson, we will build a mental map of the AWS data landscape.

---

1. Kingdom 1: AWS AI Services (The "Pre-Trained")

These are for Developers and Business Users who don't want to know about math. These services are "API-driven"—you simply send your data (image, audio, text) to an endpoint, and AWS returns the answer.

• The Philosophy: "I want the result, I don't want to train the model."
• Key Services: Amazon Rekognition, Comprehend, Polly, Transcribe, Lex.
• Analogy: Ordering a Pizza. You don't need to know how to bake; you just call the number and the pizza arrives.

---

2. Kingdom 2: AWS Machine Learning (The "SageMaker" Realm)

This is for Data Scientists and ML Engineers. This is where you go when the pre-trained services aren't enough, and you need to build something unique to your business.

• The Philosophy: "I have unique data and I need a custom model to win."
• Key Services: Amazon SageMaker (and all its sub-tools like Ground Truth, Studio, and Canvas).
• Analogy: Cooking a Gourmet Meal. You buy the raw ingredients, you use the specialized kitchen tools (SageMaker), and you follow your own recipe.

---

3. Kingdom 3: AWS Analytics (The "Foundation")

AI and ML cannot exist without data. Analytics services are the "Plumbing" that moves information into the AI/ML engines.

• The Philosophy: "I need to store, clean, and visualize my data so I can then do AI."
• Key Services: Amazon S3 (Storage), Amazon Glue (Cleaning), Amazon Athena (Querying), Amazon QuickSight (Visualizing).
• Analogy: The Grocery Store. This is where the ingredients come from and how they are organized.

---

4. The Decision Matrix

Requirement	Kingdom	Primary Service
"I need to read text from a photo."	AI Service	Amazon Textract
"I need to build a custom model for a self-driving boat."	ML Service	Amazon SageMaker
"I need to group 10 years of sales data into a chart."	Analytics	Amazon QuickSight
"I need to access a Foundation Model like Claude."	GenAI (AI)	Amazon Bedrock

graph TD
    subgraph Data_Storage
    A[Amazon S3]
    end
    
    subgraph Analytics_Plumbing
    B[AWS Glue: Clean]
    C[Amazon Athena: Query]
    end
    
    subgraph Intelligence_Consumption
    D[AI Services: API Access]
    E[SageMaker: Custom Model Build]
    F[Bedrock: GenAI API]
    end
    
    A --> B
    B --> C
    C --> D
    C --> E
    C --> F

---

5. Summary: Follow the Value

• Analytics tells you What happened? (Historical).
• AI/ML tells you What will happen? or What is this? (Inference).
• Generative AI tells you Can you create this for me? (Creation).

---

Exercise: Categorize the Request

A logistics company has 500 million sensor logs from their trucks stored in an S3 bucket. They want to create a dashboard to visualize the "Avg Fuel Consumption per City." Which kingdom should they look at?

• A. AI Services.
• B. Generative AI.
• C. Machine Learning (SageMaker).
• D. Analytics.

The Answer is D! Visualizing historical data on a dashboard is an Analytics task (likely using Amazon QuickSight). You don't need a prediction or a creation; you just need to see the facts.

---

Knowledge Check

What's Next?

Now that we have the map, how do we choose between "AI" and "ML"? It’s the classic "Managed vs. Custom" debate. Find out in Lesson 2: When to use managed AI services vs custom ML.