Amazon AWS AI Practitioner AIF-C01 GenAI Use Cases And Foundation Model Lifecycle Practice Test

 

AIF-C01 skills 2.1 | 28 original questions

This AWS Certified AI Practitioner AIF-C01 practice test focuses on genai use cases and foundation model lifecycle through original scenario-based questions aligned to AWS Exam Guide version 1.1 published April 30, 2026. Use the full ExamSnap AIF-C01 collection for broader practice across all five current exam domains. For broader exam preparation, review the Amazon AWS Certified AI Practitioner AIF-C01 Exam Dumps page.

Instructions: Select the best answer for each question. Review the rationale after answering. Each distractor includes a brief explanation of why it is not the strongest fit for the stated scenario.

Question 1

During a design review for Humongous Insurance, the risk manager must converse with customers and draft or perform support actions within controlled workflows. The team also wants to reduce manual handling. What should the team choose? The recommendation must solve the stated requirement without introducing unrelated platform complexity. The project has 5 downstream consumers and a monthly review of approximately 827 sampled interactions.

1. Customer service agent
2. Code generation
3. Summarization
4. Speech recognition
5. Recommendation explanation

Correct answer: A

Why: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. It directly addresses the requirement in this scenario.

Option review:

A: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. It directly addresses the requirement in this scenario.

B: Code generation is a common GenAI developer use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Speech recognition maps audio speech signals to text. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: GenAI can augment recommendation experiences with natural-language synthesis. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Customer service agent – GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately.

Question 2

Graphic Design Institute is moving a claims-processing redesign from pilot to production. The key decision is how to collect production observations and user signals that inform evaluation and future improvements. Which option is the strongest fit if the team wants to use current managed AWS capabilities? The design must remain supportable after launch, but no additional feature is required beyond the stated need. The rollout spans 2 application teams, each using the same approved requirement set for the next 864 days.

1. Data selection
2. Feedback
3. Model selection
4. Deployment
5. Model evaluation

Correct answer: B

Why: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. It directly addresses the requirement in this scenario.

Option review:

A: Data selection influences model capability, risk, coverage, and legal/compliance posture. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. It directly addresses the requirement in this scenario.

C: Model selection chooses the best fit before customization or deployment. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Deployment operationalizes the model for use. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Feedback – Feedback closes the lifecycle loop and can reveal issues not seen in offline tests.

Question 3

A workshop at Relecloud focuses on a single decision: how to generate personalized descriptions or ranked suggestions using context about users and items. Which option should the risk manager recommend? A short pilot window means the team prefers an approach that can be evaluated with clear success criteria. The evaluation set contains examples from 7 business workflows and 901 recent production cases.

1. Video or audio generation
2. Summarization
3. Recommendation explanation
4. Computer vision
5. Agentic AI workflow

Correct answer: C

Why: GenAI can augment recommendation experiences with natural-language synthesis. It directly addresses the requirement in this scenario.

Option review:

A: Generative models can produce audio and video as well as text and images. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: GenAI can augment recommendation experiences with natural-language synthesis. It directly addresses the requirement in this scenario.

D: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Agentic AI is suitable when the system must reason about steps, use tools, and act toward a goal. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Recommendation explanation – GenAI can augment recommendation experiences with natural-language synthesis.

Question 4

For the developer-productivity pilot at Adventure Works Manufacturing, stakeholders need to compare candidate foundation models against modality, quality, latency, cost, and governance requirements. Which concept, service, or technique most directly addresses this goal? The architecture board will reject a choice that addresses a different problem from the one described. The initial rollout covers 938 internal users across 4 business units.

1. Model evaluation
2. Model training
3. Monitoring and retraining
4. Model selection
5. Pre-training

Correct answer: D

Why: Model selection chooses the best fit before customization or deployment. It directly addresses the requirement in this scenario.

Option review:

A: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Training optimizes model parameters from examples. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Monitoring detects drift or degradation; retraining refreshes the model when justified. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Model selection chooses the best fit before customization or deployment. It directly addresses the requirement in this scenario.

E: Pre-training establishes the general capabilities of a foundation model. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Model selection – Model selection chooses the best fit before customization or deployment.

Question 5

Proseware Services is comparing alternatives for its fraud-review pilot. The risk manager needs to create new media clips or synthesized audio from instructions. Which option is most appropriate while trying to meet a strict latency target? Budget has been approved for the project, but the team still wants to avoid unnecessary recurring consumption. The workload processes about 975 requests during its busiest hour and has a documented fallback path.

1. Computer vision
2. Recommendation explanation
3. AI assistant
4. Fraud detection
5. Video or audio generation

Correct answer: E

Why: Generative models can produce audio and video as well as text and images. It directly addresses the requirement in this scenario.

Option review:

A: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI can augment recommendation experiences with natural-language synthesis. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Fraud detection commonly uses ML to identify transactions that differ from expected patterns. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Generative models can produce audio and video as well as text and images. It directly addresses the requirement in this scenario.

Learning point: Video or audio generation – Generative models can produce audio and video as well as text and images.

Question 6

An architecture review at Lucerne Publishing has narrowed a analytics modernization decision to one requirement: measure model quality, safety, and business fit before and after changes. What should the security architect select? The team will validate the result with representative production examples before rollout. The pilot uses 52 representative records from 6 approved data sources.

1. Evaluation
2. Feedback
3. Monitoring and retraining
4. Model training
5. Data collection

Correct answer: A

Why: Evaluation is required to compare models and validate readiness. It directly addresses the requirement in this scenario.

Option review:

A: Evaluation is required to compare models and validate readiness. It directly addresses the requirement in this scenario.

B: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Monitoring detects drift or degradation; retraining refreshes the model when justified. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Training optimizes model parameters from examples. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Evaluation – Evaluation is required to compare models and validate readiness.

Question 7

The risk manager at Lamna Healthcare is preparing a recommendation for a compliance-assistant prototype. The recommendation must produce translated text while preserving meaning and context. Which choice is the best match? The pilot has representative data, and the team will measure the selected approach against an agreed acceptance threshold. The first release supports 3 departments and is reviewed every 89 days.

1. AI assistant
2. Translation
3. Natural language processing
4. Video or audio generation
5. Summarization

Correct answer: B

Why: GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs. It directly addresses the requirement in this scenario.

Option review:

A: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs. It directly addresses the requirement in this scenario.

C: NLP analyzes and transforms human language. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Generative models can produce audio and video as well as text and images. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Translation – GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs.

Question 8

Contoso Retail has completed discovery for a forecasting initiative. Before implementation, the security architect must decide how to learn broad model parameters from large-scale data before task-specific adaptation. Which choice best satisfies that requirement? The team will document the rationale for auditors and wants the recommendation to be defensible from the scenario facts. The service has a 126-millisecond internal response target for the affected workflow.

1. Feedback
2. Monitoring and retraining
3. Pre-training
4. Data collection
5. Fine-tuning

Correct answer: C

Why: Pre-training establishes the general capabilities of a foundation model. It directly addresses the requirement in this scenario.

Option review:

A: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Monitoring detects drift or degradation; retraining refreshes the model when justified. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Pre-training establishes the general capabilities of a foundation model. It directly addresses the requirement in this scenario.

D: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Fine-tuning changes model weights to better fit a target behavior or domain. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Pre-training – Pre-training establishes the general capabilities of a foundation model.

Question 9

While planning a customer-support modernization, Fourth Coffee identifies this requirement: create new marketing artwork from a text description. Which option should the risk manager prioritize if the goal is to limit exposure of sensitive data? The team wants the least complex technically correct choice that satisfies the requirement. The team is comparing 5 candidate designs after a 163-day proof of concept.

1. AI assistant
2. Customer service agent
3. Summarization
4. Image generation
5. Speech recognition

Correct answer: D

Why: GenAI can synthesize new images from prompts or other conditioning inputs. It directly addresses the requirement in this scenario.

Option review:

A: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: GenAI can synthesize new images from prompts or other conditioning inputs. It directly addresses the requirement in this scenario.

E: Speech recognition maps audio speech signals to text. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Image generation – GenAI can synthesize new images from prompts or other conditioning inputs.

Question 10

A proof of concept at Margie Travel exposed a design decision for the security architect: the solution must adapt a pretrained model with additional task- or domain-specific examples. Which option most directly solves that problem? The workload has passed basic feasibility checks, so the remaining question is which approach best matches the requirement. The control owner requires evidence from 2 test groups before the 200-day release review.

1. Monitoring and retraining
2. Data preprocessing
3. Model evaluation
4. Feedback
5. Fine-tuning

Correct answer: E

Why: Fine-tuning changes model weights to better fit a target behavior or domain. It directly addresses the requirement in this scenario.

Option review:

A: Monitoring detects drift or degradation; retraining refreshes the model when justified. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Preprocessing converts raw data into a usable and consistent form. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Fine-tuning changes model weights to better fit a target behavior or domain. It directly addresses the requirement in this scenario.

Learning point: Fine-tuning – Fine-tuning changes model weights to better fit a target behavior or domain.

Question 11

School of Fine Art is documenting the target state for a contact-center transformation. The risk manager needs a solution that can answer natural-language questions by synthesizing retrieved information. Which option is the strongest fit? Stakeholders have ruled out a broad redesign and want the choice that most precisely addresses the stated need. The project has 7 downstream consumers and a monthly review of approximately 237 sampled interactions.

1. Generative search
2. Recommendation system
3. Knowledge base with retrieval
4. Translation
5. Summarization

Correct answer: A

Why: GenAI can combine retrieval and generation to improve search interactions. It directly addresses the requirement in this scenario.

Option review:

A: GenAI can combine retrieval and generation to improve search interactions. It directly addresses the requirement in this scenario.

B: Recommendation systems predict or rank items that a user may prefer. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: A knowledge base supports retrieval of relevant organizational information for search or grounded GenAI. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Generative search – GenAI can combine retrieval and generation to improve search interactions.

Question 12

Northwind Analytics is reviewing a operations automation program. The security architect has one primary requirement: make an accepted model available to applications through an inference mechanism. Which choice best fits the requirement? Operational ownership is already assigned, so the team is comparing technical fit rather than staffing models. The rollout spans 4 application teams, each using the same approved requirement set for the next 274 days.

1. Pre-training
2. Deployment
3. Model training
4. Data preprocessing
5. Data collection

Correct answer: B

Why: Deployment operationalizes the model for use. It directly addresses the requirement in this scenario.

Option review:

A: Pre-training establishes the general capabilities of a foundation model. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Deployment operationalizes the model for use. It directly addresses the requirement in this scenario.

C: Training optimizes model parameters from examples. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Preprocessing converts raw data into a usable and consistent form. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Deployment – Deployment operationalizes the model for use.

Question 13

During a design review for Litware Financial, the risk manager must condense a long document into a shorter representation of its main points. The team also wants to reduce manual handling. What should the team choose? Existing application interfaces can accommodate any of the listed choices, so functional fit is the deciding factor. The evaluation set contains examples from 9 business workflows and 311 recent production cases.

1. Customer service agent
2. Speech recognition
3. Summarization
4. AI assistant
5. Computer vision

Correct answer: C

Why: Summarization is a common LLM use case. It directly addresses the requirement in this scenario.

Option review:

A: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Speech recognition maps audio speech signals to text. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Summarization is a common LLM use case. It directly addresses the requirement in this scenario.

D: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Summarization – Summarization is a common LLM use case.

Question 14

1. Datum Research is moving a internal search upgrade from pilot to production. The key decision is how to choose representative and permitted data before pre-training or customization. Which option is the strongest fit if the team wants to use current managed AWS capabilities? Assume the required AWS capabilities are available in the selected Region and normal governance controls are in place. The initial rollout covers 348 internal users across 6 business units.
2. Model selection
3. Data collection
4. Data preprocessing
5. Data selection
6. Feedback

Correct answer: D

Why: Data selection influences model capability, risk, coverage, and legal/compliance posture. It directly addresses the requirement in this scenario.

Option review:

A: Model selection chooses the best fit before customization or deployment. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Preprocessing converts raw data into a usable and consistent form. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Data selection influences model capability, risk, coverage, and legal/compliance posture. It directly addresses the requirement in this scenario.

E: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Data selection – Data selection influences model capability, risk, coverage, and legal/compliance posture.

Question 15

A workshop at Coho Winery focuses on a single decision: how to draft or transform source code from natural-language or code context. Which option should the risk manager recommend? The review committee wants a direct mapping from the requirement to the chosen capability. The workload processes about 385 requests during its busiest hour and has a documented fallback path.

1. Fraud detection
2. Speech recognition
3. Computer vision
4. Knowledge base with retrieval
5. Code generation

Correct answer: E

Why: Code generation is a common GenAI developer use case. It directly addresses the requirement in this scenario.

Option review:

A: Fraud detection commonly uses ML to identify transactions that differ from expected patterns. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Speech recognition maps audio speech signals to text. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: A knowledge base supports retrieval of relevant organizational information for search or grounded GenAI. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Code generation is a common GenAI developer use case. It directly addresses the requirement in this scenario.

Learning point: Code generation – Code generation is a common GenAI developer use case.

Question 16

For the knowledge-assistant rollout at Lucerne Retail, stakeholders need to collect production observations and user signals that inform evaluation and future improvements. Which concept, service, or technique most directly addresses this goal? The solution will serve multiple internal teams, so the recommendation should be reusable without changing the core requirement. The pilot uses 422 representative records from 8 approved data sources.

1. Feedback
2. Model training
3. Data collection
4. Deployment
5. Data selection

Correct answer: A

Why: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. It directly addresses the requirement in this scenario.

Option review:

A: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. It directly addresses the requirement in this scenario.

B: Training optimizes model parameters from examples. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Deployment operationalizes the model for use. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Data selection influences model capability, risk, coverage, and legal/compliance posture. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Feedback – Feedback closes the lifecycle loop and can reveal issues not seen in offline tests.

Question 17

Tailspin Toys is comparing alternatives for its claims-processing redesign. The risk manager needs to support conversational questions, drafting, synthesis, and task guidance. Which option is most appropriate while trying to meet a strict latency target? The decision must follow the workload characteristics rather than a preference for the largest model or newest service. The first release supports 5 departments and is reviewed every 459 days.

1. Speech recognition
2. AI assistant
3. Customer service agent
4. Forecasting
5. Image generation

Correct answer: B

Why: GenAI assistants use natural-language interaction to help users perform knowledge work. It directly addresses the requirement in this scenario.

Option review:

A: Speech recognition maps audio speech signals to text. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI assistants use natural-language interaction to help users perform knowledge work. It directly addresses the requirement in this scenario.

C: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Forecasting predicts future values, often from time-series data. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: GenAI can synthesize new images from prompts or other conditioning inputs. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: AI assistant – GenAI assistants use natural-language interaction to help users perform knowledge work.

Question 18

An architecture review at City Power and Light has narrowed a personalization program decision to one requirement: learn broad model parameters from large-scale data before task-specific adaptation. What should the security architect select? The security baseline is already defined; the decision here concerns the specific capability described in the requirement. The service has a 496-millisecond internal response target for the affected workflow.

1. Evaluation
2. Model selection
3. Pre-training
4. Data selection
5. Model evaluation

Correct answer: C

Why: Pre-training establishes the general capabilities of a foundation model. It directly addresses the requirement in this scenario.

Option review:

A: Evaluation is required to compare models and validate readiness. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Model selection chooses the best fit before customization or deployment. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Pre-training establishes the general capabilities of a foundation model. It directly addresses the requirement in this scenario.

D: Data selection influences model capability, risk, coverage, and legal/compliance posture. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Pre-training – Pre-training establishes the general capabilities of a foundation model.

Question 19

The risk manager at Consolidated Messenger is preparing a recommendation for a developer-productivity pilot. The recommendation must converse with customers and draft or perform support actions within controlled workflows. Which choice is the best match? The recommendation must solve the stated requirement without introducing unrelated platform complexity. The team is comparing 7 candidate designs after a 533-day proof of concept.

1. Knowledge base with retrieval
2. Generative search
3. Computer vision
4. Customer service agent
5. AI assistant

Correct answer: D

Why: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. It directly addresses the requirement in this scenario.

Option review:

A: A knowledge base supports retrieval of relevant organizational information for search or grounded GenAI. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI can combine retrieval and generation to improve search interactions. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. It directly addresses the requirement in this scenario.

E: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Customer service agent – GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately.

Question 20

Nod Publishers has completed discovery for a fraud-review pilot. Before implementation, the security architect must decide how to choose representative and permitted data before pre-training or customization. Which choice best satisfies that requirement? The design must remain supportable after launch, but no additional feature is required beyond the stated need. The control owner requires evidence from 4 test groups before the 570-day release review.

1. Evaluation
2. Data preprocessing
3. Model training
4. Model evaluation
5. Data selection

Correct answer: E

Why: Data selection influences model capability, risk, coverage, and legal/compliance posture. It directly addresses the requirement in this scenario.

Option review:

A: Evaluation is required to compare models and validate readiness. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Preprocessing converts raw data into a usable and consistent form. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Training optimizes model parameters from examples. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Data selection influences model capability, risk, coverage, and legal/compliance posture. It directly addresses the requirement in this scenario.

Learning point: Data selection – Data selection influences model capability, risk, coverage, and legal/compliance posture.

Question 21

While planning a analytics modernization, Fabrikam Health identifies this requirement: create new marketing artwork from a text description. Which option should the risk manager prioritize if the goal is to limit exposure of sensitive data? A short pilot window means the team prefers an approach that can be evaluated with clear success criteria. The project has 9 downstream consumers and a monthly review of approximately 607 sampled interactions.

1. Image generation
2. Speech recognition
3. Recommendation system
4. Summarization
5. Customer service agent

Correct answer: A

Why: GenAI can synthesize new images from prompts or other conditioning inputs. It directly addresses the requirement in this scenario.

Option review:

A: GenAI can synthesize new images from prompts or other conditioning inputs. It directly addresses the requirement in this scenario.

B: Speech recognition maps audio speech signals to text. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Recommendation systems predict or rank items that a user may prefer. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Image generation – GenAI can synthesize new images from prompts or other conditioning inputs.

Question 22

A proof of concept at Wingtip Logistics exposed a design decision for the security architect: the solution must compare candidate foundation models against modality, quality, latency, cost, and governance requirements. Which option most directly solves that problem? The architecture board will reject a choice that addresses a different problem from the one described. The rollout spans 6 application teams, each using the same approved requirement set for the next 644 days.

1. Model evaluation
2. Model selection
3. Evaluation
4. Data selection
5. Fine-tuning

Correct answer: B

Why: Model selection chooses the best fit before customization or deployment. It directly addresses the requirement in this scenario.

Option review:

A: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Model selection chooses the best fit before customization or deployment. It directly addresses the requirement in this scenario.

C: Evaluation is required to compare models and validate readiness. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Data selection influences model capability, risk, coverage, and legal/compliance posture. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Fine-tuning changes model weights to better fit a target behavior or domain. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Model selection – Model selection chooses the best fit before customization or deployment.

Question 23

Trey Research is documenting the target state for a forecasting initiative. The risk manager needs a solution that can produce translated text while preserving meaning and context. Which option is the strongest fit? Budget has been approved for the project, but the team still wants to avoid unnecessary recurring consumption. The evaluation set contains examples from 3 business workflows and 681 recent production cases.

1. Customer service agent
2. AI assistant
3. Translation
4. Summarization
5. Video or audio generation

Correct answer: C

Why: GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs. It directly addresses the requirement in this scenario.

Option review:

A: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs. It directly addresses the requirement in this scenario.

D: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Generative models can produce audio and video as well as text and images. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Translation – GenAI models can perform translation, although managed translation services may be preferred for predictable specialized needs.

Question 24

Bellows College is reviewing a customer-support modernization. The security architect has one primary requirement: make an accepted model available to applications through an inference mechanism. Which choice best fits the requirement? The team will validate the result with representative production examples before rollout. The initial rollout covers 718 internal users across 8 business units.

1. Model evaluation
2. Pre-training
3. Monitoring and retraining
4. Deployment
5. Data collection

Correct answer: D

Why: Deployment operationalizes the model for use. It directly addresses the requirement in this scenario.

Option review:

A: Evaluation assesses whether the model meets technical and business acceptance criteria. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Pre-training establishes the general capabilities of a foundation model. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Monitoring detects drift or degradation; retraining refreshes the model when justified. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Deployment operationalizes the model for use. It directly addresses the requirement in this scenario.

E: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Deployment – Deployment operationalizes the model for use.

Question 25

During a design review for Blue Yonder Airlines, the risk manager must create new media clips or synthesized audio from instructions. The team also wants to reduce manual handling. What should the team choose? The pilot has representative data, and the team will measure the selected approach against an agreed acceptance threshold. The workload processes about 755 requests during its busiest hour and has a documented fallback path.

1. Computer vision
2. Summarization
3. Code generation
4. AI assistant
5. Video or audio generation

Correct answer: E

Why: Generative models can produce audio and video as well as text and images. It directly addresses the requirement in this scenario.

Option review:

A: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Summarization is a common LLM use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Code generation is a common GenAI developer use case. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: GenAI assistants use natural-language interaction to help users perform knowledge work. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Generative models can produce audio and video as well as text and images. It directly addresses the requirement in this scenario.

Learning point: Video or audio generation – Generative models can produce audio and video as well as text and images.

Question 26

Woodgrove Bank is moving a contact-center transformation from pilot to production. The key decision is how to adapt a pretrained model with additional task- or domain-specific examples. Which option is the strongest fit if the team wants to use current managed AWS capabilities? The team will document the rationale for auditors and wants the recommendation to be defensible from the scenario facts. The pilot uses 792 representative records from 2 approved data sources.

1. Fine-tuning
2. Data selection
3. Evaluation
4. Monitoring and retraining
5. Feedback

Correct answer: A

Why: Fine-tuning changes model weights to better fit a target behavior or domain. It directly addresses the requirement in this scenario.

Option review:

A: Fine-tuning changes model weights to better fit a target behavior or domain. It directly addresses the requirement in this scenario.

B: Data selection influences model capability, risk, coverage, and legal/compliance posture. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Evaluation is required to compare models and validate readiness. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: Monitoring detects drift or degradation; retraining refreshes the model when justified. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Feedback closes the lifecycle loop and can reveal issues not seen in offline tests. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Fine-tuning – Fine-tuning changes model weights to better fit a target behavior or domain.

Question 27

A workshop at Wide World Importers focuses on a single decision: how to generate personalized descriptions or ranked suggestions using context about users and items. Which option should the risk manager recommend? The team wants the least complex technically correct choice that satisfies the requirement. The first release supports 7 departments and is reviewed every 829 days.

1. Customer service agent
2. Recommendation explanation
3. Computer vision
4. Generative search
5. Agentic AI workflow

Correct answer: B

Why: GenAI can augment recommendation experiences with natural-language synthesis. It directly addresses the requirement in this scenario.

Option review:

A: GenAI can power support agents when grounding, permissions, and escalation paths are designed appropriately. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: GenAI can augment recommendation experiences with natural-language synthesis. It directly addresses the requirement in this scenario.

C: Computer vision is designed to extract information from images or video. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

D: GenAI can combine retrieval and generation to improve search interactions. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Agentic AI is suitable when the system must reason about steps, use tools, and act toward a goal. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Recommendation explanation – GenAI can augment recommendation experiences with natural-language synthesis.

Question 28

For the sales-assistant rollout at VanArsdel Media, stakeholders need to measure model quality, safety, and business fit before and after changes. Which concept, service, or technique most directly addresses this goal? The workload has passed basic feasibility checks, so the remaining question is which approach best matches the requirement. The service has a 866-millisecond internal response target for the affected workflow.

1. Data selection
2. Pre-training
3. Evaluation
4. Data collection
5. Model selection

Correct answer: C

Why: Evaluation is required to compare models and validate readiness. It directly addresses the requirement in this scenario.

Option review:

A: Data selection influences model capability, risk, coverage, and legal/compliance posture. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

B: Pre-training establishes the general capabilities of a foundation model. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

C: Evaluation is required to compare models and validate readiness. It directly addresses the requirement in this scenario.

D: Data collection establishes the source material used by downstream preparation and training. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

E: Model selection chooses the best fit before customization or deployment. This can be appropriate in another scenario, but it does not most directly satisfy the requirement described here.

Learning point: Evaluation – Evaluation is required to compare models and validate readiness.