AI Workforce Final Assessment

Hands-on checklistI completed every required module.I selected a real low-risk business workflow.I classified and minimized the data.I confirmed the approved tool or method.I included human review and permission checks.I defined a pilot, baseline and success measures.I defined rollback, stop and incident-reporting paths.I reviewed the submission for restricted information.

Sanitized evidence

Never enter passwords, access tokens, payment-card data, health information, confidential client records or other restricted information.

Business problem and processUse a sanitized summary. Do not enter credentials or restricted records.

Data classification and controlsUse a sanitized summary. Do not enter credentials or restricted records.

Approved tool selectionUse a sanitized summary. Do not enter credentials or restricted records.

Structured sanitized promptUse a sanitized summary. Do not enter credentials or restricted records.

Human review and risk controlsUse a sanitized summary. Do not enter credentials or restricted records.

Pilot and measurement planUse a sanitized summary. Do not enter credentials or restricted records.

Knowledge check

Minimum required: 12 correct out of 15.

1. What type of workflow should be used for the capstone?

An irreversible financial approval.A restricted personnel decision.Password management.A real low-risk workflow.

2. What must happen before data is used?

Credential sharing.Removal of ownership.Automatic upload.Classification, approval and minimization.

3. What makes a tool selection defensible?

The longest feature list.A vendor slogan.Popularity alone.Task fit, data fit, controls, permissions, licensing and fallback.

4. What should a complete prompt include?

Only a topic.Goal, context, constraints, output format and verification plan.A secret key.A request to guess missing facts.

5. What should happen before consequential action?

Deletion of records.Human review and approval.Automatic execution.Permission expansion.

6. How should an unsupported claim be handled?

Verify, qualify or remove it.Repeat it.Publish it.Treat it as a source.

7. What is prompt injection?

A style guide.Untrusted instructions attempting to redirect the system.A spreadsheet formula.A login method.

8. Why review permissions?

To increase automation.To remove audit records.To improve font size.To limit source and output access to authorized people and systems.

9. What is the purpose of a pilot?

To test safety, quality and value in a controlled scope.To launch everywhere.To skip training.To remove the fallback.

10. What should baseline measures represent?

A vendor estimate only.The current process before implementation.A guaranteed future result.A hidden target.

11. What should happen when a stop condition occurs?

Remove the owner.Pause, escalate and use fallback or rollback.Hide the event.Scale immediately.

12. What should a change log record?

Only the date.Version, reason, owner, testing and approval.Passwords.Nothing.

13. Who owns the final decision?

The browser.The accountable human decision-maker.No one.The AI model.

14. What evidence should be submitted?

Passwords.Sanitized summaries and approved records.Restricted source files.Access tokens.

15. What does completion prove?

Demonstrated safe and practical application under the course standard.Permission to ignore company policy.Expertise in every AI product.Guaranteed business results.

I used sanitized evidence, followed approved-account and minimum-necessary-data rules, verified important outputs and kept final decision ownership with a person.

Learn, apply, protect, verify and document.

What you will practice

Define the business problem and process

Classify and minimize data

Select an approved tool and fallback

Provide a structured sanitized prompt

Define human review and risk controls

Define pilot and measurement plan