Skip to content

GoAI 101 Practice Center

Eight judgment challenges: AI-use judgment, workflow design, data and context, tool permissions, governance, testing, procurement requirements, and acceptance decisions.

Info

Every scenario is simulated teaching material; agency names, people, and figures are examples only and do not represent any real institution.

Warning

Do not enter real national ID numbers, real citizen personal data, sensitive government information, account passwords, API keys, or other unauthorized data into these exercises. Use de-identified, simulated, or public data only.

Practice challenges

Activity 1: Should AI Do This?

Pick or receive a government task scenario, judge what AI may and may not do, and record the initial risk.

Steps

  1. Write down the business goal and the user
  2. List what AI may and may not do
  3. Note the human owner and the initial risk
  4. Judge whether it is worth designing further

Deliverables

  • A draft Government AI Use Case Card

Activity 2: From Task to Government Agentic Workflow

Break Activity 1's task into a workflow: sequence, routing, AI decision points, and human decision points.

Steps

  1. List tasks and their inputs/outputs
  2. Draw sequential/parallel steps and routing
  3. Mark AI decision points and human decision points
  4. Mark escalation and the failure path

Deliverables

  • A draft Government Agentic Workflow Canvas

Activity 3: What Should the AI Know?

For the chosen use case, classify candidate information into Include / Retrieve / Remember / Exclude.

Steps

  1. List candidate information
  2. Classify as Include / Retrieve / Remember / Exclude
  3. Mark the source and sensitivity
  4. Write the exclusion reason

Deliverables

  • A draft Government Context & Data Canvas

Activity 4: Can the Agent Use This Tool?

List the tools or systems the workflow uses, and classify each action as Read / Write / Recommend / Execute / Requires Approval / Prohibited.

Steps

  1. List the tools and systems
  2. Mark the data access scope
  3. Classify each action
  4. Mark the owner and fallback

Deliverables

  • A draft Government AI Tool & Permission Matrix

Activity 5: Where Must a Human Be in the Loop?

Mark each workflow step as AI only / AI + verification / human decision / human approval / prohibited automation.

Steps

  1. List every workflow step
  2. Mark the control level
  3. Mark approval gates and control gates
  4. List the matching risk

Deliverables

  • A draft Government AI Governance Checklist

Activity 6: Test the AI, Not the Demo

Write at least three test cases for the chosen use case: normal, edge/ambiguous, and failure/restricted.

Steps

  1. Choose a use case
  2. Write at least three test cases
  3. Define expected behavior and forbidden behavior
  4. Record evidence and pass/fail

Deliverables

  • At least three Government AI Test Case entries

Activity 7: From Use Case to Procurement Requirement

Turn one workflow requirement into a procurement requirement, defining a verification method and an acceptance criterion.

Steps

  1. Pick a workflow requirement
  2. Write it as a concrete requirement
  3. Define the verification method
  4. Define the acceptance criterion and required evidence

Deliverables

  • A draft Government AI Procurement Requirements Sheet (at least one requirement)

Activity 8: Would You Accept This System?

Check a simulated acceptance evidence pack or vendor demo against the 14 acceptance dimensions, make an acceptance decision, and cite evidence.

Steps

  1. Check the 14 acceptance dimensions
  2. Find the matching evidence
  3. Make an Accept / Accept with conditions / Reject decision
  4. State the reasoning and open issues

Deliverables

  • A draft Government AI Procurement & Acceptance Checklist

Completion marks are stored only in this browser.

Sync my progressJoin a session to keep progress across devices and share it with your instructor

Info

After you join, your completed lesson numbers are sent to your instructor together with your learner code, so they can see your progress. Use the codes your instructor gave you, and do not enter your name, company secrets or other personal data. Progress is kept with the rest of that session's data.

What next?