Module 01 · Fundamentals & literacy
Classical ML &
Generative AI.
Learn where prediction, creation, and clear rules each belong—and make the business decision before the technology decision.
Choose Play when you are ready.
Classical ML & Generative AIChoose Play when you are ready.
English captions available.
What you will take away
Choose the right job for AI.
- Match a business task to prediction, generative content, or an explicit rule.
- Separate a forecast from a verified fact or authorized commitment.
- Compare quality, review effort, operating cost, and business value.
A question to keep in mind
What must be predicted, what must be created, and what must be verified?Try the scenario near the end of the lesson before listening to the answer.
Transcript and captions
Read the lesson
Download the complete transcriptEnglish captions
English captions · SRTThe full lesson uses approximately 125 MB per narrator. Video loads as you watch; playback does not start automatically. Captions follow speech chunks and are not word-aligned.
Keep the decision boundary clear.
| Business need | Useful approach | What to verify |
|---|---|---|
| Estimate a delivery risk | Classical machine learning | Relevant historical data, prediction quality and the cost of missed alerts. |
| Draft a customer update | Generative AI grounded in approved records | Factual support, current information and human approval of commitments. |
| Apply an eligibility policy | Explicit rules | Clear conditions, authoritative policy and accountable ownership. |
Practise the business decision.
Harborline Distribution needs earlier warnings about late deliveries and clearer customer updates. It has reliable delivery history, current operational records and a rule that account managers must approve customer commitments. Which approach best fits?
- Use one generative model for every task and measure employee satisfaction.
- Predict delivery risk, ground message drafts in verified records and retain human approval for customer commitments.
- Treat the predicted risk as a confirmed delivery date and send it automatically.
- Fine-tune on historical emails and automatically send messages that match the company’s tone.
Reveal the answer and compare the alternatives
B. The prediction identifies risk; current records support the draft; the account manager retains authority to commit. Each component is evaluated for the job it performs.
A does not establish forecast reliability or factual accuracy. C turns an estimate into a promise. D may reproduce a tone, but historical emails do not establish current delivery facts or authority to send a commitment.
Original educational practice, not an official AWS exam question. This is the first published lesson; the complete 40-lesson course is still in production.