← All courses

🧭 Free course · 12 lessons · Start here

AI at Work Foundations

What AI tools can and cannot do, when to trust them, and how to get useful answers on the first try.

about 65 minutes total, one short lesson at a timefor office workers, managers, non-technical professionals

Your first win: Write a useful context-task-format prompt and know when to verify AI output.

1. What AI actually is (and is not)

An LLM predicts plausible text from patterns; it is not a database of facts

2. Your first real prompt: context, task, format

A useful prompt states the context, the task, and the output format

3. When to trust the answer

Verify anything the model recalled on its own; trust transformations of material you supplied

4. Give it a role

Assigning the model a role focuses its expertise and tone on your situation

5. Show, don't tell: examples beat descriptions

One or two examples of the output you want beat paragraphs of description

6. The critique loop: make it improve its own work

Asking the model to critique and revise its own output reliably raises quality

7. Ask better follow-up questions

The first answer is a starting point; targeted follow-ups that name what to change beat starting over

8. Make AI ask you questions first

When your request is fuzzy, tell the AI to ask you questions before it answers

9. Use your own facts first

Supply the facts yourself and ask the model to do the language work, not the recall

10. Check before you send

Run a quick verification pass over numbers, claims, and tone before forwarding AI-assisted work

11. Private, sensitive, or safe to paste?

Classify information as public, internal, or confidential before putting it into AI

12. Your personal AI task map

Choose where to use AI by mapping your tasks on value/frequency against risk