Seeing Around Corners

AI isn't just about "Talking" or "Seeing." It is a mathematical engine for Prediction. In business, the two most valuable predictions are:

1. What is going to happen? (Forecasting).
2. Is someone trying to cheat us? (Fraud Detection).

In our final lesson of Module 8, we look at the specialized services that handle these complex time-series and behavioral patterns.

---

1. Amazon Forecast (The Fortune Teller)

Amazon Forecast is a fully managed service that uses machine learning to deliver highly accurate forecasts. It uses the same technology that Amazon.com uses to predict demand for millions of items.

Why use Forecast instead of Excel?

Traditional forecasting (like in a spreadsheet) just looks at your past sales and draws a line. Amazon Forecast can look at the "Big Picture":

• Seasonality: Knowing that people buy more coats in October.
• Holidays: Adjusting for the chaos of Black Friday.
• Related Variables: Knowing that "Rainy weather" increases the sale of umbrellas.

Use Case: A grocery store predicting how much milk they will need next Tuesday to avoid overstocking (waste) or understocking (lost sales).

---

2. Amazon Fraud Detector (The Digital Guard)

Amazon Fraud Detector is a fully managed service that makes it easy to identify potentially fraudulent online activities, such as online payment fraud and the creation of fake accounts.

How it works:

You provide historical data of "Safe" and "Fraudulent" activities. When a new event happens (e.g., a $500 purchase from a new IP address), Fraud Detector evaluates it against your model and gives it a Risk Score.

Use Case: An e-commerce site blocking a transaction because it "matches the pattern" of a known credit card theft ring.

---

3. The "Decision Support" Relationship

graph TD
    subgraph Data_Inputs
    A[Historical Sales / Logs]
    B[External Events: Weather/Holidays]
    end
    
    subgraph Analysis_Engines
    C[Amazon Forecast]
    D[Amazon Fraud Detector]
    end
    
    subgraph Business_Impact
    E[Inventory Optimization]
    F[Financial Security]
    G[Better Cashflow]
    end
    
    A & B --> C & D
    C --> E & G
    D --> F

---

4. Summary Table: Business Decision Services

Service	Best For...	Input Data
Amazon Forecast	Demand, Inventory, Revenue	Time-series data (Date + Values)
Fraud Detector	Payments, Account Signups	Event data (IP, Email, Total)
Amazon Rekognition	Content Moderation	Images / Videos
Amazon Personalize	Recommendations	User Clickstreams

---

5. Recap of Module 8

We have applied AI across the whole corporation:

• Support: Using Lex and Connect to automate conversations.
• Marketing: Using Personalize and Bedrock to make every customer feel unique.
• Ops: Using Kendra and Amazon Q to find and summarize company knowledge.
• Finance/Strategy: Using Forecast and Fraud Detector to protect the bottom line.

---

Knowledge Check

---

What's Next?

We’ve learned the services and the use cases. But how do we decide when to "Buy" a service and when to "Build" one? In Module 9: AI Adoption Strategy, we look at the high-level business decisions that define a successful AI project.