# Practical AI skills workbook

Universal AI Institute — Research. Education. Equity.

Use fictional or shareable public data. Keep the prompts, outputs, and your own review notes.

## Summarize without losing the meaning

run a two-pass review
1. Run the extraction prompt and compare every field with the notes.
2. Underline any item that changes the strength of the source: proposed becomes approved, may becomes will, or discussed becomes decided.
3. Correct the extraction before asking for the final brief.
4. Ask for a revised brief based only on the corrected extraction.
5. Check that every sentence in the brief can be traced back to the notes.

**Your finished artifact:** The original notes, corrected extraction, and final brief. A teammate should be able to identify both the next action and what remains unresolved.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next:

## Write with AI. Keep your voice.

revise with a reason
1. Write a 120-word paragraph about something you personally learned. Use only details you are comfortable sharing.
2. State the point you want the reader to understand.
3. Ask for feedback before requesting a rewrite.
4. Accept one suggested change, modify another, and reject a third. If there are fewer than three suggestions, ask for additional options.
5. Write a one-sentence reason for each decision.
6. Read the final paragraph aloud. Replace phrases you would not naturally use.

**Your finished artifact:** Your original paragraph, a revision, and a short editorial decision log. You should be able to explain every factual statement and every important change.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next:

## From numbers to a defensible answer

build a checked comparison
1. Calculate both attendance rates yourself.
2. Calculate both positive response rates among respondents.
3. Ask the assistant for the same metrics and compare all four results.
4. Ask it to explain whether Session B’s larger positive-response count means higher satisfaction. Check the explanation against the denominators.
5. Write a three-sentence brief: one result, one limitation, and one follow-up question.

**Your finished artifact:** A calculation sheet and a brief. A strong brief might note the different survey response rates and ask whether the sessions served comparable audiences.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next:

## Build your first AI-powered workflow

test the process three times
1. Prepare three short fictional note sets. Include a clear action, an action with no deadline, and a proposal that has not been approved.
2. Run the same instruction on each set.
3. Review each output against the original notes. Count invented fields, omitted actions, and incorrectly labeled proposals.
4. Record generation time, review time, and correction time. Compare the total with doing the task yourself.
5. Revise the instruction if needed and rerun all three cases, including the one that originally worked.

**Your finished artifact:** A one-page workflow specification, three reviewed outputs, and a short keep/revise/stop decision. Include the evidence for that decision.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next: