Build an AIF-C01 Exam Decision Map

Create a requirement-first decision map and weighted AIF-C01 diagnostic.

Introduction

30 Second Summary

Certification questions can make several familiar services look equally plausible. A single overlooked requirement can turn a confident guess into the wrong answer.

In this project, you will build AIF-C01-readiness.txt, a requirement-first decision map spanning the five official AIF-C01 exam domains. You will test it with a 10-question weighted sprint that reveals your two highest-priority study areas.

What You'll Build

After the sprint, your saved scorecard will point straight to the two domains that deserve your next study session.

By the end of this project, you'll have:

  • A reusable five-domain scorecard that exposes the keywords influencing your first answer choices.
  • A decision ladder you can use to choose classification, regression, clustering, prompt engineering, RAG, fine-tuning, or a custom foundation model. Its control matrix keeps access, sensitive-data discovery, activity logging, compliance documents, and generative AI safeguards separate.
  • A weighted diagnostic that turns ten scenario answers into five domain scores. The result identifies two evidence-based study priorities. Your follow-on field points to the free AWS Official Practice Question Set on AWS Skill Builder.
  • Secret Mission: Two rewritten scenarios that prove one decisive requirement can change the correct answer.

Are there any prerequisites?

You'll need a Windows device with basic AWS knowledge. Your existing AWS Builder ID is enough because no paid AWS Skill Builder subscription is required.

Before We Start

Before any hands-on work, take a moment to commit to the reusable AIF-C01 exam decision map you are about to build. Naming its purpose now gives every later study choice a clear target.

Set Up Your Free Exam Workspace

Exam preparation gets messy when source pages disappear into old browser tabs. Scattered notes make every study session start from zero.

A reliable browser keeps the current official material close. Microsoft Edge gives this Windows workspace a consistent starting point.

AWS Skill Builder holds the free official practice set. Notepad keeps your scorecard in one file on the Desktop.

In this step, get ready to:
  • Prepare Microsoft Edge with the official exam pages.
  • Confirm access to the Official Practice Question Set.
  • Save your AIF-C01 readiness scorecard on the Desktop.
Check Microsoft Edge and open the official sources

The exam guide defines the knowledge covered by AIF-C01. The certification preparation page points you to official practice activities.

  • Press the Windows key to open Start search.
  • Type Microsoft Edge into Start search.

✔️ Microsoft Edge is available

  • Select Microsoft Edge from the search results.

Microsoft Edge is ready to hold your official exam references.

ⓧ Microsoft Edge is missing

  • Open the official Microsoft Edge download page in another available browser.
  • Follow the installer prompts to add Microsoft Edge to Windows.
  • Press the Windows key when the installation finishes.
  • Type Microsoft Edge into Start search.
  • Select Microsoft Edge from the search results.

Microsoft Edge is now available for the rest of the project.

Microsoft Edge still not opening?

  • Restart Windows if the installer finishes without adding Microsoft Edge to Start search.
  • Run the official installer again if the installation stops before completion.

Help me troubleshoot Microsoft Edge installation on Windows.

Your browser now holds the current exam guide plus the official path to practice materials. Those two tabs give every later decision a dependable source.

Why official sources instead of an AWS lab?

The AIF-C01 target candidate selects appropriate technologies for business requirements. Coding models or deploying pipelines falls outside that scope.

An AWS Management Console lab would add deployment work without improving this decision map. This project creates no AWS account resources.

Verify AWS Skill Builder access

The free Official Practice Question Set contains 20 questions with detailed answer feedback. This project confirms that you can reach it without starting the full set.

  • Open AWS Skill Builder in a new Microsoft Edge tab.
  • Select Sign In.
  • Choose AWS Builder ID as the sign-in method.
  • Use your existing AWS Builder ID to finish signing in.
  • Open the Course Catalog.
  • Set the Training Category filter to Exam Preparation.
  • Search for AWS Certified AI Practitioner.
  • Confirm the Official Practice Question Set is listed without starting its questions.

That secures your official follow-on activity. The free 20-question set is ready for a separate study session.

Cannot reach the question set?

  • Confirm that AWS Skill Builder shows your signed-in session before opening the Course Catalog.
  • Clear the existing search text before searching for AWS Certified AI Practitioner again.
  • Confirm that the Training Category filter is set to Exam Preparation.

Help me locate the Official Practice Question Set in AWS Skill Builder.

Create the scorecard in Notepad

A local text file gives every practice result the same home. Its sections separate your baseline from the decision map you build later.

  • Press the Windows key to open Start search.
  • Type Notepad into Start search.

✔️ Notepad is available

  • Select Notepad from the search results.

You should see a blank Notepad page ready for the scorecard.

ⓧ Notepad is missing

  • Open the Windows Notepad listing.
  • Follow the Microsoft Store prompts to install Windows Notepad.
  • Press the Windows key when installation finishes.
  • Type Notepad into Start search.
  • Select Notepad from the search results.

You should now see a blank Notepad page.

Notepad still missing?

  • Restart Windows if the Microsoft Store finishes installing Notepad without adding it to Start search.
  • Return to the Windows Notepad listing if the installation did not finish.

Help me troubleshoot Windows Notepad installation.

  • Build the scorecard by copying this template into the blank Notepad page:
AIF-C01 READINESS SCORECARD
Date:

BASELINE
Domain 1:
Domain 2:
Domain 3:
Domain 4:
Domain 5:
Keyword trap:

DECISION MAP
ML problem type:
FM customization:
Prompt technique:
Responsible AI:
Security and governance:

WEIGHTED SPRINT
Domain 1 /20:
Domain 2 /24:
Domain 3 /28:
Domain 4 /14:
Domain 5 /14:
Project score /100:

STUDY PRIORITIES
1.
2.
Official follow-on:

How is the scorecard organized?

  • BASELINE records your first answers plus the keyword trap behind them.
  • DECISION MAP holds the requirement-first rules you develop.
  • WEIGHTED SPRINT reserves points that align with the five official content-domain percentages.
  • STUDY PRIORITIES turns your final result into two focused follow-on targets.
  • Select File in Notepad.
  • Select Save As.

You should see the Save As window with controls for the file location plus its name.

  • Select Desktop as the save location.
  • Enter AIF-C01-readiness.txt as the file name.

The save location should now point to your Desktop. The filename should include the .txt extension.

  • Select Save.

✔️ Awesome, I've got everything!

  • Confirm Notepad shows the complete scorecard with Official follow-on: as its final line.

Your saved scorecard is ready for the baseline exercise.

ⓧ I'd like to double check the full code

  • Compare Desktop\AIF-C01-readiness.txt with this complete reference.
AIF-C01 READINESS SCORECARD
Date:

BASELINE
Domain 1:
Domain 2:
Domain 3:
Domain 4:
Domain 5:
Keyword trap:

DECISION MAP
ML problem type:
FM customization:
Prompt technique:
Responsible AI:
Security and governance:

WEIGHTED SPRINT
Domain 1 /20:
Domain 2 /24:
Domain 3 /28:
Domain 4 /14:
Domain 5 /14:
Project score /100:

STUDY PRIORITIES
1.
2.
Official follow-on:

What should match?

Every heading plus score line should appear in the same order. Blank lines keep the four sections easy to scan during timed practice.

Before you close Notepad, do you expect the scorecard to reopen with every heading intact?

  • Close Notepad.
  • Double-click AIF-C01-readiness.txt on the Desktop.

Your scorecard should reopen in Notepad with the complete template intact. That proves the workspace has a durable home.

  • Return to Microsoft Edge from earlier.
  • Select the AIF-C01 exam guide tab.

You should see the current guide at https://docs.aws.amazon.com/aws-certification/latest/ai-practitioner-01/ai-practitioner-01.html.

  • Select the AWS Certification Preparation tab.

You should see the certification preparation page at https://aws.amazon.com/certification/certification-prep/.

  • Select the AWS Skill Builder tab.

You should still be signed in. The Course Catalog should retain your AWS Certified AI Practitioner search results.

Scorecard did not reopen correctly?

  • Check the Desktop for AIF-C01-readiness.txt if double-clicking another file opens different text.
  • Return to Notepad if the scorecard is missing any section.
  • Use the full-file comparison above to restore any missing lines.

Help me fix my saved AIF-C01 readiness scorecard.

Your official sources plus reusable scorecard are ready. Next, you will test what happens when familiar keywords drive your exam answers.

Expose the Keyword Trap

Your scorecard is ready in Notepad. You now have one place to capture how you reason through exam scenarios.

Several AIF-C01 options share familiar AWS words. A keyword-first choice can feel convincing before the full requirement is tested. This baseline deliberately tests that habit under time pressure.

In this step, get ready to:
  • Answer five scenarios using your first instinct.
  • Compare your responses with the requirement-based answer key.
  • Record the keyword trap as a reusable decision rule.
Answer five rapid scenarios

Your first responses create an honest baseline. The goal is to capture your current reasoning before the answer key can influence it.

Before you start, which scenario do you expect to be hardest to answer from one familiar keyword?

  • Switch back to Desktop\AIF-C01-readiness.txt in Notepad.
  • Find the BASELINE section.
  • Give yourself four minutes for the five scenarios.

Four-Minute Scenario Pack

  • Scenario A, Domain 1. A retailer has historical transactions labeled fraudulent or legitimate. It must assign each new transaction to one of those categories. Choose from classification / regression / clustering / generative AI.
  • Scenario B, Domains 2 and 3. An employee assistant must answer from frequently changing internal manuals. It must reference the source. It must avoid retraining the foundation model. Choose from few-shot prompting / RAG / fine-tuning / custom pre-training.
  • Scenario C, Domain 3. A team places three paired input and desired-output examples in a prompt. The model parameters stay unchanged. Choose from zero-shot prompting / few-shot prompting / RAG / fine-tuning.
  • Scenario D, Domain 4. A generative AI application needs configurable safeguards for safety. It also needs safeguards for privacy. It also needs safeguards for veracity. Choose from Amazon Bedrock Guardrails / AWS CloudTrail / IAM / AWS Artifact.
  • Scenario E, Domain 5. A team needs to control who can invoke its AI system. It needs to record who performed API actions. It needs to download compliance reports. Choose only one service from IAM / AWS CloudTrail / AWS Artifact.
  • Record the five responses in the existing Domain lines. Put Scenario A on Domain 1:. Put Scenario B on both Domain 2: and Domain 3:. Put Scenario C on Domain 3:. Put Scenario D on Domain 4:. Put Scenario E on Domain 5:.
  • Capture the first keyword that influenced every response. Scenario A: first keyword that influenced Scenario A. Scenario B: first keyword that influenced Scenario B. Scenario C: first keyword that influenced Scenario C. Scenario D: first keyword that influenced Scenario D. Scenario E: first keyword that influenced Scenario E.
  • Save Desktop\AIF-C01-readiness.txt by pressing Ctrl+S.

You now have an honest record of your first instinct. That baseline preserves the reasoning habit you are about to test.

Check the baseline answers

The answer key checks the requirement behind each scenario. Scenario E provides the crucial test because its prompt demands one service for several different needs.

  • Compare your five responses with the checked Domain lines below:
Domain 1: Scenario A: Classification.
Domain 2: Scenario B: RAG.
Domain 3: Scenario B: RAG. Scenario C: Few-shot prompting.
Domain 4: Scenario D: Amazon Bedrock Guardrails.
Domain 5: Scenario E: AMBIGUOUS. IAM maps to fine-grained access. AWS CloudTrail maps to user and API activity. AWS Artifact maps to security and compliance documents.

Why These Answers Fit

  • Scenario A uses labeled categories. Classification predicts categorical labels.
  • Scenario B needs frequently changing external knowledge without retraining the core model. RAG retrieves that knowledge when the assistant answers.
  • Scenario C provides multiple paired examples inside the prompt. That structure is few-shot prompting.
  • Scenario D asks for configurable safeguards on generative AI inputs and outputs. Amazon Bedrock Guardrails provides those safeguards.
  • Scenario E combines three different controls. IAM handles fine-grained access. AWS CloudTrail records user and API activity. AWS Artifact provides security and compliance documents.

Scenario E's forced single-service choice is the intended shortfall. No single listed service covers all three requirements.

  • Count one point for each matching response across Scenarios A through D.
  • Count no point for Scenario E.
  • Update the five Domain lines in Desktop\AIF-C01-readiness.txt to match the checked version above.
  • Save Desktop\AIF-C01-readiness.txt by pressing Ctrl+S.
  • Review the BASELINE section. You should see four checked answers plus Scenario E marked AMBIGUOUS.

Unsure Why an Answer Changed?

Compare the scenario's required output with the reason listed in the answer callout. Focus on labels, knowledge freshness, prompt examples, safeguards, or control type.

Ask for help tracing the deciding requirement in one of my baseline scenarios.

Record the limitation

A useful baseline records the clue that pulled your attention. Pairing that clue with a decision rule turns the mistake into something you can catch during future practice.

  • Replace the existing Keyword trap: line plus the blank area before DECISION MAP with this completed block:
Keyword trap:
Scenario A influence: [[SCENARIO_A_INFLUENCE="first keyword that influenced Scenario A"]]
Scenario B influence: [[SCENARIO_B_INFLUENCE="first keyword that influenced Scenario B"]]
Scenario C influence: [[SCENARIO_C_INFLUENCE="first keyword that influenced Scenario C"]]
Scenario D influence: [[SCENARIO_D_INFLUENCE="first keyword that influenced Scenario D"]]
Scenario E influence: [[SCENARIO_E_INFLUENCE="first keyword that influenced Scenario E"]]
Rule: Decompose the requirement before selecting the service.

What Does This Rule Change?

The influence lines expose which familiar words pulled you toward an answer. The rule tells you to separate the required outcomes before choosing a technique or service.

Scenario E becomes three smaller decisions about access, activity records, and compliance documents. Each decision maps to its own service.

  • Save Desktop\AIF-C01-readiness.txt by pressing Ctrl+S.
  • Confirm that every Scenario influence line contains the keyword you captured.

Missing an Influence Word?

Use the first familiar word that pulled your attention before you checked the answer key. Choose the word that influenced your original response.

Ask for help identifying which keyword influenced my original choice.

Before you reopen the file, do you expect the completed baseline to preserve both your original influences and the corrected reasoning?

  • Close Notepad.
  • Double-click AIF-C01-readiness.txt on your Desktop to reopen it.
  • Review the reopened BASELINE section.

You should see Classification for Scenario A. You should also see RAG for Scenario B plus Few-shot prompting for Scenario C.

You should see Amazon Bedrock Guardrails for Scenario D. Scenario E should show AMBIGUOUS with access, activity logging, and compliance documents mapped to separate services.

That is the trap exposed. Your scorecard now preserves the exact reasoning gap that the decision map needs to fix.

✔️ Awesome, I've got everything!

Your completed baseline is saved. Keep Desktop\AIF-C01-readiness.txt open in Notepad for the next step.

ⓧ I'd like to double check the full code

Compare your scorecard with the complete file below. Your five influence words should reflect your own first responses.

AIF-C01 READINESS SCORECARD
Date:

BASELINE
Domain 1: Scenario A: Classification.
Domain 2: Scenario B: RAG.
Domain 3: Scenario B: RAG. Scenario C: Few-shot prompting.
Domain 4: Scenario D: Amazon Bedrock Guardrails.
Domain 5: Scenario E: AMBIGUOUS. IAM maps to fine-grained access. AWS CloudTrail maps to user and API activity. AWS Artifact maps to security and compliance documents.
Keyword trap:
Scenario A influence: [[SCENARIO_A_INFLUENCE="first keyword that influenced Scenario A"]]
Scenario B influence: [[SCENARIO_B_INFLUENCE="first keyword that influenced Scenario B"]]
Scenario C influence: [[SCENARIO_C_INFLUENCE="first keyword that influenced Scenario C"]]
Scenario D influence: [[SCENARIO_D_INFLUENCE="first keyword that influenced Scenario D"]]
Scenario E influence: [[SCENARIO_E_INFLUENCE="first keyword that influenced Scenario E"]]
Rule: Decompose the requirement before selecting the service.

DECISION MAP
ML problem type:
FM customization:
Prompt technique:
Responsible AI:
Security and governance:

WEIGHTED SPRINT
Domain 1 /20:
Domain 2 /24:
Domain 3 /28:
Domain 4 /14:
Domain 5 /14:
Project score /100:

STUDY PRIORITIES
1.
2.
Official follow-on:

Your baseline now shows why keyword matching falls short. Next, you will build the requirement-first decision map that resolves the ambiguity.

Build a Requirement-First Decision Map

Your completed baseline exposed the gap in keyword matching. Scenario E bundled three different needs into one prompt.

A requirement-first decision map gives each need its own branch. In this step, you will turn those branches into a reusable map in Notepad.

In this step, get ready to:
  • Distinguish problem types through requirement signals.
  • Distinguish customization choices through requirement signals.
  • Map exam cues to the matching techniques or controls.
Add the problem-type and customization ladders

The problem-type ladder starts with labels before checking the required output. The customization ladder starts with the simplest option that can meet the requirement.

  • In Notepad, scroll to the DECISION MAP section.
  • Select the empty fields from ML problem type: through FM customization:.
  • Fill both ladders by replacing the selected fields with this text:
ML problem type:
- Labeled data plus categorical output: classification.
- Labeled data plus continuous numeric output: regression.
- Unlabeled data plus a need to discover groups or patterns: clustering.
- Generated text, images, audio, summaries, or assistants: consider generative AI and foundation models.
FM customization:
- Start with prompt engineering when clearer instructions can meet the requirement.
- Use RAG when the answer must use external or frequently changing knowledge without retraining the core model.
- Consider fine-tuning or continued pre-training when better task performance, labeled domain data, or domain-specific language is required.
- Consider a custom foundation model only when available pre-trained models do not meet the requirement, substantial proprietary data is available, and full control is needed.

How do these ladders work?

  • Classification fits a categorical output. Regression fits a continuous numeric output.
  • Clustering fits unlabeled data when the goal is to discover groups.
  • Generative AI fits requests for generated content. Foundation models support this branch.
  • Prompt engineering is the first customization branch when clearer instructions can solve the task.
  • RAG adds external knowledge without retraining the core model.
  • Fine-tuning uses relevant labeled data to improve task performance. Continued pre-training can strengthen domain-specific language.
  • A custom foundation model fits requirements that need substantial proprietary data plus full control.
  • Save your changes to AIF-C01-readiness.txt.
  • Review the two new ladders in Notepad.

You should see four problem-type branches followed by four customization branches.

Branches in the wrong section?

Check that ML problem type: appears directly above its four branches. Check that FM customization: appears directly above its four branches.

If the original headings disappeared, restore them from the text block above. Help me place the problem-type and customization ladders in my scorecard.

Add prompt and evaluation cues

Prompt questions often hinge on whether examples appear inside the prompt. Evaluation questions require both model performance signals and business outcomes.

  • Select the empty fields from Prompt technique: through Responsible AI:.
  • Fill the selected fields by copying this text:
Prompt technique:
- Zero-shot: an instruction with no examples.
- Few-shot: multiple paired input and desired-output examples inside the prompt.
- Evaluate technical performance and business value using accuracy, precision, recall, F1 score, task completion rate, user satisfaction, and cost per interaction.
Responsible AI:
- Check bias, fairness, inclusivity, robustness, safety, and veracity.
- Use representative and balanced datasets, label-quality analysis, human audits, and subgroup analysis.
- Use Amazon Bedrock Guardrails when the requirement is configurable generative AI safeguards for safety, privacy, or veracity.

What do these cues separate?

  • Zero-shot prompting uses an instruction without examples.
  • Few-shot prompting includes multiple paired examples inside the prompt.
  • Technical metrics measure model performance. Business metrics measure the result for users or the organization.
  • Responsible AI checks how an AI system behaves across people plus conditions.
  • Amazon Bedrock Guardrails maps to configurable safeguards for generative AI inputs or outputs.
  • Save your changes to AIF-C01-readiness.txt.
  • Review the prompt and responsible AI branches in Notepad.

You should see two prompt patterns followed by evaluation cues plus three responsible AI branches.

Prompt or safeguard cues missing?

Confirm that the evaluation line sits under Prompt technique:. Confirm that all three safeguard branches sit under Responsible AI:.

Help me compare my prompt and responsible AI branches with the required decision map.

Add the security and governance control matrix

Security questions can combine several control categories in one scenario. A control matrix gives each requirement a separate destination.

  • Select the empty Security and governance: field.
  • Build the control matrix by replacing that field with this text:
Security and governance:
- Fine-grained access to AWS workload resources: IAM.
- Sensitive-data discovery: Amazon Macie.
- User activity and API usage records: AWS CloudTrail.
- On-demand AWS security and compliance documents: AWS Artifact.
- AI output grounding, validation, and hallucination reduction: RAG grounding, output validation, and confidence scoring.

How does the control matrix work?

  • IAM maps to fine-grained access for AWS workload resources.
  • Amazon Macie maps to sensitive-data discovery.
  • AWS CloudTrail maps to user activity plus API usage records.
  • AWS Artifact maps to AWS security plus compliance documents.
  • RAG grounding connects an answer to retrieved knowledge.
  • Output validation checks generated results before they are used.
  • Confidence scoring helps express uncertainty in an AI output.
  • Save your changes to AIF-C01-readiness.txt.
  • Review the five mappings under Security and governance:.

You should see one requirement category on the left side of every mapping. You should see its matching control on the right side.

Control matrix looks incomplete?

Check for five mappings under the security and governance heading. The first four map AWS requirements to services.

The final mapping covers grounding plus validation controls. Help me find a missing control-matrix mapping.

Before you check Scenario E, which control should match each of its three requirements?

  • Return to Scenario E in the BASELINE section.
  • Match its access requirement against the completed control matrix.
  • Match its activity-recording requirement against the completed control matrix.
  • Match its compliance-document requirement against the completed control matrix.

You should map access to IAM. You should map activity records to AWS CloudTrail. You should map compliance documents to AWS Artifact.

  • Record the resolved answer below the control matrix by copying this line:
Scenario E resolution: IAM + AWS CloudTrail + AWS Artifact.

Why does the answer use three services?

Scenario E contains three independent requirements. Each service satisfies one requirement from the control matrix.

Decomposition turns the ambiguous single-choice prompt into a defensible combined answer.

  • Save your changes to AIF-C01-readiness.txt.
  • Review every branch in the completed DECISION MAP section.

You should see at least one requirement signal paired with a matching choice in every branch. The final line should show IAM + AWS CloudTrail + AWS Artifact as the Scenario E resolution.

That closes the gap your baseline exposed. Each clause now lands on its own control.

Scenario E still feels ambiguous?

Read one clause at a time. Match each clause to one row in the security and governance matrix.

Help me decompose Scenario E into its three requirements.

✔️ Awesome, I've got everything!

Your decision map is complete. Keep AIF-C01-readiness.txt saved on your Desktop for the weighted sprint.

ⓧ I'd like to double check the full code

Compare your completed DECISION MAP section with this reference. Keep your completed baseline plus the later scorecard fields unchanged.

ML problem type:
- Labeled data plus categorical output: classification.
- Labeled data plus continuous numeric output: regression.
- Unlabeled data plus a need to discover groups or patterns: clustering.
- Generated text, images, audio, summaries, or assistants: consider generative AI and foundation models.
FM customization:
- Start with prompt engineering when clearer instructions can meet the requirement.
- Use RAG when the answer must use external or frequently changing knowledge without retraining the core model.
- Consider fine-tuning or continued pre-training when better task performance, labeled domain data, or domain-specific language is required.
- Consider a custom foundation model only when available pre-trained models do not meet the requirement, substantial proprietary data is available, and full control is needed.
Prompt technique:
- Zero-shot: an instruction with no examples.
- Few-shot: multiple paired input and desired-output examples inside the prompt.
- Evaluate technical performance and business value using accuracy, precision, recall, F1 score, task completion rate, user satisfaction, and cost per interaction.
Responsible AI:
- Check bias, fairness, inclusivity, robustness, safety, and veracity.
- Use representative and balanced datasets, label-quality analysis, human audits, and subgroup analysis.
- Use Amazon Bedrock Guardrails when the requirement is configurable generative AI safeguards for safety, privacy, or veracity.
Security and governance:
- Fine-grained access to AWS workload resources: IAM.
- Sensitive-data discovery: Amazon Macie.
- User activity and API usage records: AWS CloudTrail.
- On-demand AWS security and compliance documents: AWS Artifact.
- AI output grounding, validation, and hallucination reduction: RAG grounding, output validation, and confidence scoring.
Scenario E resolution: IAM + AWS CloudTrail + AWS Artifact.

Your decision map now turns requirements into defensible choices. Next, you will test it under time pressure in the weighted scenario sprint.

Run the Weighted Scenario Sprint

Your requirement-first decision map has turned the ambiguous baseline into clear choices. The next challenge is applying those rules while the clock is running.

This sprint tests your reasoning under time pressure. You will then convert every mistake into a domain-specific study priority.

In this step, get ready to:
  • Complete a timed 10-question scenario pack.
  • Calculate a project score using the official domain weights.
  • Turn your two lowest-performing domains into a focused study plan.
Complete the 10-question pack in 10 minutes

A short timer tests whether your decision rules are easy to retrieve. Your first-pass answers also give you honest evidence for the study plan.

  • Switch back to the Notepad file from earlier.
  • Place your cursor directly above Domain 1 /20: under WEIGHTED SPRINT.
  • Add ten answer lines labeled Q1: through Q10:.

You should now see a place for every answer above the five domain score lines.

  • Set a 10-minute timer on a device available to you.
  • Record one choice for Questions 1 through 9 as you work through the pack below.
  • Record four requirement-to-service matches for Question 10.

Questions 1 to 5

  • Question 1. Domain 1: A company has labeled historical support tickets. It must predict one of four predefined issue categories for each new ticket. Which technique fits? A. Regression. B. Classification. C. Clustering. D. RAG.
  • Question 2. Domain 1: A retailer has unlabeled customer behavior data. It wants to discover naturally similar customer groups. Which technique fits? A. Classification. B. Regression. C. Clustering. D. Few-shot prompting.
  • Question 3. Domain 2: The same valid prompt can produce different outputs across invocations. Which GenAI limitation best describes this? A. Inclusivity. B. Nondeterminism. C. Clustering. D. Data residency.
  • Question 4. Domain 2. Multiple response: Which factors are in scope when selecting a GenAI model? A. Compliance. B. Office color scheme. C. Cost. D. Latency. E. Model complexity.
  • Question 5. Domain 3: An assistant must answer questions from frequently updated company manuals. It must include source references. Model retraining must be avoided. Which approach fits? A. Custom pre-training. B. RAG. C. Clustering. D. Regression.

Questions 6 to 10

  • Question 6. Domain 3: Prompt engineering plus RAG did not meet a specialized task's accuracy requirement. The company has relevant labeled data plus domain-specific terminology. Which next approach fits? A. Zero-shot prompting. B. Clustering. C. Fine-tuning. D. AWS Artifact.
  • Question 7. Domain 3: A prompt contains an instruction with no demonstration examples. Which technique is this? A. Zero-shot prompting. B. Few-shot prompting. C. RAG. D. Fine-tuning.
  • Question 8. Domain 3: A prompt contains three paired examples of inputs plus desired outputs before a new input. Which technique is this? A. Regression. B. RAG. C. Zero-shot prompting. D. Few-shot prompting.
  • Question 9. Domain 4: A company needs configurable safeguards for safety, privacy, plus veracity on generative AI inputs plus outputs. Which service fits? A. AWS Artifact. B. Amazon Bedrock Guardrails. C. AWS CloudTrail. D. Amazon Macie.
  • Question 10. Domain 5. Matching: Match each requirement to one service. Requirements: a. fine-grained workload access. b. sensitive-data discovery. c. user plus API activity records. d. security plus compliance documents. Services: IAM, Amazon Macie, AWS CloudTrail, AWS Artifact.
  • Stop your timer after Question 10.
  • Check that every answer line contains your first-pass choice.

You have completed the timed sprint. Those ten answers now capture how your decision map performs under pressure.

Score by official domain weight

The project weighting mirrors the five official AIF-C01 domain percentages. Each domain therefore contributes to the total according to its share of the scored exam content.

Answer Key and Project Weighting

  • Domain 1: Question 1 is B. Classification. Question 2 is C. Clustering. Award 10 points for each correct answer. The maximum is 20 points.
  • Domain 2: Question 3 is B. Nondeterminism. Question 4 is A, C, D, plus E. These choices represent Compliance, Cost, Latency, plus Model complexity only. Award 12 points for Question 3. Award 12 points for Question 4 only when every correct choice is selected plus the distractor is excluded. The maximum is 24 points.
  • Domain 3: Question 5 is B. RAG. Question 6 is C. Fine-tuning. Question 7 is A. Zero-shot prompting. Question 8 is D. Few-shot prompting. Award 7 points for each correct answer. The maximum is 28 points.
  • Domain 4: Question 9 is B. Amazon Bedrock Guardrails. Award 14 points for the correct answer.
  • Domain 5: Question 10 matches a to IAM. Match b to Amazon Macie. Match c to AWS CloudTrail. Match d to AWS Artifact. Award 14 points only when every pair is correct.
  • Compare your ten recorded answers with the answer key.
  • Award points using the rule for each domain.
  • Record your five results beside Domain 1 /20:, Domain 2 /24:, Domain 3 /28:, Domain 4 /14:, plus Domain 5 /14:.

Your scorecard now shows how many weighted points each domain contributed.

  • Add the five domain scores to calculate your result out of 100.
  • Record the sum beside Project score /100:.

The five domain maxima total 100. Your recorded project score should match the sum of the five domain lines.

  • Add the note Formative project score. This is not an official exam score or pass prediction. below your project score.
  • Save Desktop\AIF-C01-readiness.txt by pressing Ctrl+S.

That gives you a weighted diagnostic with clear boundaries. It measures this project sprint without making a claim about your exam outcome.

Turn the result into a study plan

The domain maxima differ because the official exam weights differ. Converting each result to a percentage lets you compare your performance across all five domains.

  • Calculate each domain percentage by dividing its score by its maximum before multiplying the result by 100.
  • Rank the five domains from the lowest percentage earned to the highest.

You should now have an ordered domain ranking based on performance instead of raw points.

  • Prioritize the domain with the larger official exam weight whenever two percentages tie.
  • Enter the two lowest-performing domains under STUDY PRIORITIES.
  • Record one missed decision-map rule beside each priority.

Your two priorities now connect a measured weakness to a specific reasoning rule you can review.

The free official set contains 20 questions with detailed answer feedback. It belongs in a separate study session so this project stays focused on building your reasoning framework.

  • Record Free AWS Official Practice Question Set, 20 questions beside Official follow-on:.
  • Return to the AWS Certification Preparation tab from earlier in Microsoft Edge.
  • Open the free Official Practice Question Set in AWS Skill Builder.
  • Select the bookmark control in the Microsoft Edge address bar to save the page for later.
  • Leave the 20 questions for a separate study session.

Before the final check, which sections do you expect to contain the clearest evidence of your strengths plus your next study targets?

  • Save Desktop\AIF-C01-readiness.txt by pressing Ctrl+S.
  • Close Notepad.
  • Double-click AIF-C01-readiness.txt on your Desktop to reopen it.
  • Confirm that the five domain scores add up to the recorded project score out of 100.
  • Confirm that STUDY PRIORITIES contains two domains plus one missed decision rule for each.
  • Confirm that Official follow-on: contains Free AWS Official Practice Question Set, 20 questions.

Strong finish. Your reopened scorecard now turns a timed result into two evidence-based study priorities plus a free official follow-on activity.

Secret mission

Flip the Requirement, Flip the Answer

One changed clause can turn a confident exam answer into the wrong one. Rewrite customization and security scenarios to prove that the decisive requirement controls the answer.

Clean Up Your Resources

Clean Up Your Resources

Your AIF-C01 study workspace created one local scorecard. This project has no deployed AWS resources or ongoing costs.

Decide whether to keep your scorecard ready for future study sessions, close your study apps for now, or delete the file entirely.

Resources you used:

  • Desktop\AIF-C01-readiness.txt, your saved Notepad scorecard.

Keep everything running

No action needed. Choose this if you plan to keep using your decision map during AIF-C01 study sessions.

  • Keep Desktop\AIF-C01-readiness.txt for future scenario practice.
  • Return to the free AWS Official Practice Question Set when you are ready for a separate study session.

Pause - I'll come back to this later

Close your study apps to clear your screen. Your scorecard remains saved on the Desktop.

  • Save the completed scorecard in Notepad.
  • Close Notepad.
  • Close Microsoft Edge.

Delete - I don't want to use this again

Remove the local scorecard when you no longer want the study artifact. Permanent deletion is final.

  • Close Desktop\AIF-C01-readiness.txt in Notepad.
  • Locate AIF-C01-readiness.txt on your Windows Desktop.
  • Select AIF-C01-readiness.txt.
  • Press the Delete key to move the file to the Windows Recycle Bin.
  • Open Recycle Bin from your Desktop.
  • Check the Recycle Bin for unrelated files before continuing.
  • Permanently delete only AIF-C01-readiness.txt from the Recycle Bin.
  • Return to the Desktop to confirm that AIF-C01-readiness.txt is no longer listed.

Nice Work!

Nice Work!

You did it! You built a reusable decision map for AWS Certified AI Practitioner (AIF-C01). Your saved AIF-C01-readiness.txt now turns exam requirements into defensible answers.

What you learned:

  • Expose the keyword-matching trap by decomposing an ambiguous security requirement before selecting an AWS service.
  • Build a requirement-first decision map that connects each business requirement to a defensible exam answer.
  • Turn a domain-weighted diagnostic into two evidence-based study priorities. Save the free Official Practice Question Set in AWS Skill Builder as your follow-on activity.
  • Demonstrate through the Secret Mission how one decisive requirement flips a customization answer from RAG to fine-tuning. Apply the same rule to IAM and AWS CloudTrail security scenarios.

Ready to quiz yourself?