AIF-C01 Practice Exam 53 — 20 Questions

Question 1

A financial services company uses Amazon Bedrock with the Anthropic Claude 3 Haiku model to answer employee questions about internal policies. The knowledge base is updated weekly, and the company wants the model to cite the exact source document and page number in its responses. Which approach should the company use to meet these requirements?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Use Amazon Bedrock Knowledge Bases with a vector store and enable citations in the RetrieveAndGenerate API call.. The company needs answers grounded in specific internal documents with citations. Amazon Bedrock Knowledge Bases performs retrieval-augmented generation by fetching relevant passages from a connected data source and can return citations that point to the source document and page. This directly satisfies the requirement, whereas fine-tuning, temperature adjustments, or Guardrails do not provide source attribution.

Answer

Increase the temperature setting in the InvokeModel API call to encourage the model to include source references.

Answer

Fine-tune the Anthropic Claude 3 Haiku model on the internal policy documents and deploy the custom model.

Answer

Use Amazon Bedrock Guardrails to filter responses and automatically append document citations.

Question 2

A media company uses Amazon Bedrock to generate personalized news summaries. They notice that summaries sometimes include details not present in the source articles. They want to reduce these hallucinations without retraining the model. Which approach should they use?

Accepted Answer

Use Retrieval Augmented Generation (RAG) by retrieving relevant passages from the source articles and including them in the prompt.. Retrieval Augmented Generation supplies the model with relevant excerpts from the source articles at inference time, so the generated summary is conditioned on actual content rather than the model's internal knowledge. This reduces fabricated details without retraining. Other options either increase randomness, rely on model size, or truncate output, none of which ground the response in the provided material.

Answer

Decrease the maximum token length in the InvokeModel request.

Answer

Switch to a larger foundation model with more parameters.

Answer

Increase the model's temperature setting to encourage more creative outputs.

Question 3

A financial services company uses Amazon Bedrock with the Anthropic Claude 3 Haiku model to answer employee questions about internal policies. The knowledge base is updated weekly, and the company wants the model to cite the exact source document and page number in its responses. Which approach should the company use to meet these requirements?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Use Retrieval Augmented Generation (RAG) by storing policy documents in an Amazon Bedrock knowledge base and retrieving relevant passages at inference time.. Retrieval Augmented Generation with an Amazon Bedrock knowledge base retrieves relevant, current policy passages and supplies them to the model as context, so answers are grounded in authoritative documents. It avoids retraining costs and keeps pace with frequent updates, directly solving the outdated or incorrect policy detail problem.

Answer

Switch to a larger foundation model with a higher parameter count.

Answer

Fine-tune the foundation model on the updated policy documents.

Answer

Increase the model's temperature setting to encourage more creative and up-to-date answers.

Question 4

A retail company wants to build an application that uses a foundation model on Amazon Bedrock to answer customer questions about product availability. They need the model to access real-time inventory data from their internal database and perform actions such as reserving an item. Which TWO capabilities should they implement to achieve this? (Choose two.)

Accepted Answer

Define an action group in the agent that maps to an API schema for the inventory and reservation operations.. Amazon Bedrock Agents orchestrate tasks by invoking action groups, which are defined with OpenAPI schemas and backed by Lambda functions. This allows the agent to query real-time inventory and perform reservations. RAG over static documents, temperature changes, and logging do not provide live data access or transactional capabilities, so they cannot satisfy the requirements.

Answer

Enable model invocation logging to capture inventory queries.

Answer

Use Retrieval Augmented Generation with a static document store containing product manuals.

Answer

Increase the model's temperature to allow more creative problem-solving.

Question 5

A financial services company has deployed a machine learning model that approves or denies loan applications in real time. The compliance team requires that any applicant who is denied must receive a meaningful explanation of the decision, and the company must be able to prove which model version and input features produced each decision for audit purposes. Which AWS service should the company use to capture the model's feature attributions and store them for each inference request?

Accepted Answer

Amazon SageMaker Clarify with online explainability enabled on the endpoint. Per-decision explainability requires a capability that computes feature attributions at inference time and persists them for audit. SageMaker Clarify online explainability does exactly this inside the endpoint, returning a SHAP-based attribution for each request, and endpoint data capture stores the request and response in Amazon S3. Together they satisfy both the applicant-facing explanation requirement and the internal audit trail obligation.

Answer

Amazon SageMaker Experiments to track the training runs

Answer

Amazon SageMaker Model Monitor with a data quality baseline

Answer

AWS CloudTrail data events on the SageMaker endpoint

Question 6

A developer is building a retrieval-augmented generation (RAG) assistant on Amazon Bedrock. The assistant must answer questions about internal policy documents that change frequently, and answers must cite the source passages. The developer wants a managed capability that handles chunking, embedding, and retrieval so the application code stays minimal. Which Amazon Bedrock feature should the developer use?

Accepted Answer

Amazon Bedrock knowledge bases. Amazon Bedrock knowledge bases are the managed RAG capability that handles ingestion, chunking, embedding, vector storage, and retrieval, enabling grounded answers with source citations while keeping application code small. Evaluation, Guardrails, and Agents address quality measurement, content safety, and task orchestration respectively, none of which deliver the required retrieval pipeline on their own.

Answer

Amazon Bedrock model evaluation

Answer

Amazon Bedrock Agents

Answer

Amazon Bedrock Guardrails

Question 7

A financial services firm uses Amazon Bedrock with Anthropic Claude to analyze earnings call transcripts. They need the model to output results in a strict JSON schema for downstream processing. The model occasionally returns prose or invalid JSON. Which Amazon Bedrock feature should they use to enforce the output structure?

Accepted Answer

Converse API with a tool definition that specifies the JSON schema as input parameters.. The Converse API's tool use capability lets you define a tool whose input schema is the desired JSON structure. The model then returns arguments conforming to that schema when it invokes the tool, providing reliable structured output. Guardrails, logging, and sampling parameters do not enforce a schema, so they cannot guarantee valid JSON for downstream processing.

Answer

Increase the top_p value to 1.0 to make the output more deterministic.

Answer

Guardrails for Amazon Bedrock with a denied topics policy.

Answer

Model invocation logging to Amazon CloudWatch.

Question 8

A media company wants to automatically generate short video captions from uploaded audio files using a foundation model on AWS. The solution should transcribe speech and then produce concise captions. Which combination of AWS services should they use?

Accepted Answer

Amazon Transcribe to convert audio to text, then Amazon Bedrock to generate captions from the transcript.. The task requires two capabilities: speech-to-text and generative text creation. Amazon Transcribe provides accurate automatic speech recognition, and Amazon Bedrock supplies foundation models that turn the transcript into concise captions. Together they form a managed, scalable pipeline that meets the scenario without custom model development.

Answer

Amazon Kinesis Data Streams to ingest audio, then Amazon Bedrock to generate captions directly from the audio stream.

Answer

Amazon Rekognition to analyze the audio, then Amazon Bedrock to generate captions.

Answer

Amazon Polly to convert audio to text, then Amazon Comprehend to generate captions.

Question 9

A developer is using Amazon Bedrock to generate summaries of news articles. They notice that the model sometimes includes information not present in the original article. Which term describes this phenomenon?

Accepted Answer

Hallucination. The phenomenon where a foundation model generates information not present in the source material is called hallucination. It is a key challenge in generative AI, especially for summarization tasks. Mitigation includes grounding the model with retrieval-augmented generation or using guardrails to filter unsupported claims.

Answer

Data leakage

Answer

Underfitting

Answer

Overfitting

Question 10

A financial services firm uses Amazon Bedrock with a foundation model to draft client emails. Compliance requires that no personally identifiable information from the prompt ever appear in the response, and that the model refuse requests for investment guarantees. The team wants a managed, configurable layer rather than custom prompt engineering alone. Which combination should they apply?

Accepted Answer

Guardrails for Amazon Bedrock with sensitive information filters and a denied topics policy.. The firm needs runtime enforcement on both input and output. Guardrails for Amazon Bedrock applies sensitive information filters that detect and mask or block PII, and a denied topics policy that makes the model decline investment-guarantee requests. These are managed, configurable controls evaluated per request, unlike prompt wording, capacity reservations, or offline evaluation, none of which prevent prohibited content from reaching clients.

Answer

A higher temperature combined with a system prompt instructing the model to be careful.

Answer

Provisioned Throughput plus a custom lexicon in the tokenizer.

Answer

Amazon Bedrock model evaluation with a toxicity metric and automatic retries.

Question 11

A developer wants to compare the output quality of several foundation models available in Amazon Bedrock for a text summarization task. They need to evaluate responses side by side using the same prompt. Which Amazon Bedrock feature should they use?

Accepted Answer

Amazon Bedrock model evaluation. Amazon Bedrock model evaluation provides both automatic and human-based evaluation workflows to compare foundation models on tasks like summarization. It lets you run the same prompt across models and review results side by side. Guardrails, Agents, and provisioned throughput serve different purposes such as content filtering, orchestration, and capacity reservation, so they cannot fulfill the comparison requirement.

Answer

Amazon Bedrock provisioned throughput

Answer

Amazon Bedrock Agents

Answer

Amazon Bedrock Guardrails

Question 12

A media company runs a daily news digest built on Amazon Bedrock with Anthropic Claude. Editors complain that summaries of long policy documents sometimes omit the final recommendations, even though the source text clearly contains them near the end. The requests currently pass only the document body and set a maximum output length of 300 tokens. Which change best addresses the truncation of the source content before the model reasons over it?

Accepted Answer

Chunk the document and apply retrieval or summarization so all sections reach the model.. The symptom points to input truncation: content beyond the context window is discarded before the model sees it, so the closing recommendations vanish. Increasing output length or changing sampling cannot recover text that never entered the prompt. Chunking the document, then summarizing or retrieving relevant chunks, ensures the full source is represented and the recommendations reach the model.

Answer

Raise the temperature setting so the model explores more of the document.

Answer

Increase the maxTokens parameter so the digest can be longer.

Answer

Switch the model invocation to streaming responses.

Question 13

A media company uses Amazon Bedrock to generate personalized news summaries for its subscribers. The model occasionally produces summaries that include outdated facts from its training data. The company wants the summaries to reflect only the most recent articles from its internal content management system (CMS). The CMS exposes a REST API that returns the latest articles. Which approach should the company take to ensure the generated summaries are grounded in the latest articles?

Accepted Answer

Use Amazon Bedrock agents with an action group that calls the CMS REST API to fetch the latest articles, and include those articles in the prompt.. The requirement is to ground generated summaries in the most recent articles from a CMS. Amazon Bedrock agents with action groups can call the CMS REST API at inference time, fetch the latest articles, and pass them as context to the foundation model. This retrieval-augmented approach ensures factual recency without retraining. Other options either alter generation parameters or involve static training methods that cannot dynamically incorporate fresh content.

Answer

Fine-tune the foundation model using Amazon Bedrock continued pre-training with a small set of recent articles.

Answer

Retrain the foundation model on the CMS articles daily using Amazon Bedrock custom model training.

Answer

Increase the model's temperature setting to encourage more creative and current outputs.

Question 14

A retail company is preparing to launch a generative AI assistant built on Amazon Bedrock that will answer customer questions using internal product documents. The governance team wants to reduce the risk of the assistant producing fabricated, misleading, or harmful outputs before launch and during operation. (Choose two.)

Accepted Answer

Run Amazon Bedrock model evaluation jobs that score responses for accuracy, toxicity, and robustness against a curated dataset.. Reducing fabricated and harmful output requires both pre-launch measurement and runtime enforcement. Bedrock model evaluation jobs quantify accuracy, toxicity, and robustness so the team can make an evidence-based launch decision, while Guardrails contextual grounding and relevance checks block or flag responses unsupported by the retrieved documents. Logging, WAF rules, and storage encryption address unrelated risks.

Answer

Enable AWS CloudTrail management event logging for all Bedrock API calls in the account.

Answer

Store all assistant conversations in Amazon S3 with default encryption and a lifecycle policy.

Answer

Attach an AWS WAF web ACL with rate-based rules to the application's public endpoint.

Question 15

A healthcare company uses Amazon Bedrock with a foundation model to generate patient education materials. They must ensure that the model does not include protected health information (PHI) in its responses, even if it appears in the prompt. Which Amazon Bedrock feature should they configure?

Accepted Answer

Guardrails for Amazon Bedrock with sensitive information filters. Guardrails for Amazon Bedrock provides sensitive information filters that can detect and block or mask PII and custom patterns in both prompts and responses. Configuring these filters prevents PHI from appearing in generated materials. Model evaluation, provisioned throughput, and Agents do not offer runtime content redaction, so they cannot meet the compliance requirement.

Answer

Amazon Bedrock Agents with an action group to query a patient database

Answer

Amazon Bedrock provisioned throughput

Answer

Amazon Bedrock model evaluation with toxicity metrics

Question 16

A retail analytics team wants an assistant that answers questions about last quarter's sales using data stored in an Amazon S3 bucket of PDF reports and CSV exports. They want the model to cite the underlying documents and avoid inventing figures. Which Amazon Bedrock capability should they use?

Accepted Answer

Knowledge Bases for Amazon Bedrock with the S3 bucket as a data source.. Grounding answers in private documents requires retrieval, not filtering or capacity. Knowledge Bases for Amazon Bedrock manages ingestion, chunking, embedding, and retrieval from the S3 source, so responses are generated from the sales reports and can include citations. Evaluation, Guardrails, and Provisioned Throughput address quality measurement, content safety, and capacity respectively, none of which supply the source data.

Answer

Guardrails for Amazon Bedrock configured with a denied topics policy.

Answer

Model evaluation jobs that score responses against a ground-truth dataset.

Answer

Provisioned Throughput purchased for the foundation model.

Question 17

A financial services company is building an application on Amazon Bedrock that generates personalized investment summaries. The compliance team requires that every generated summary includes an exact, verifiable citation from the company's approved regulatory documents, and that the model must not fabricate any citation. The company has a large corpus of approved PDF documents stored in Amazon S3. Which approach should the company use to meet these requirements?

Accepted Answer

Use Amazon Bedrock Knowledge Bases to ingest the S3 documents and configure the model to generate responses grounded in retrieved passages, then verify the returned source citations.. Grounding the model in an approved document corpus is the reliable way to produce verifiable citations. Amazon Bedrock Knowledge Bases ingests the S3 documents, retrieves relevant passages per query, and returns source attributions, so generated summaries reference real approved text instead of relying on parametric memory. Sampling settings, fine-tuning, and logging do not provide retrieval-backed citation integrity.

Answer

Enable model invocation logging in Amazon Bedrock and audit the logs to confirm that any citation the model produces exists in the S3 corpus.

Answer

Increase the model's temperature setting so the model explores more of its training data and is more likely to recall the exact regulatory text.

Answer

Fine-tune the foundation model on the approved regulatory documents using Amazon Bedrock custom models, then rely on the fine-tuned model to reproduce exact citations from memory.

Question 18

A retail company wants to compare the output quality of several foundation models available in Amazon Bedrock for a product description generation task. They need a repeatable, automated way to score responses against reference descriptions. Which AWS capability should they use?

Accepted Answer

Amazon Bedrock model evaluation jobs with automatic metrics such as BERTScore and ROUGE.. Amazon Bedrock model evaluation jobs run automated scoring against reference datasets using metrics such as BERTScore and ROUGE, enabling consistent comparison of multiple foundation models. This directly supports selecting the best model for product description generation with repeatable, quantitative results rather than subjective manual review.

Answer

AWS Trusted Advisor to assess foundation model performance.

Answer

Amazon Bedrock provisioned throughput to benchmark model latency.

Answer

Amazon CloudWatch Logs to capture model responses and manually review them.

Question 19

A media company uses Amazon Bedrock to generate article summaries. They notice that for long articles, the model sometimes ignores instructions placed at the beginning of the prompt. The company wants to improve the model's adherence to instructions without changing the model or increasing cost significantly. Which prompt engineering technique should they apply?

Accepted Answer

Move the instructions to the end of the prompt and repeat them after the article text.. The model's tendency to overlook early instructions in long contexts is a known limitation. By moving instructions to the end of the prompt and repeating them after the article, the company places them where the model's attention is strongest. This prompt engineering technique improves adherence without retraining or increasing cost, making it the most effective solution.

Answer

Use a higher temperature value to make the model more creative in following instructions.

Answer

Split the article into smaller chunks and summarize each chunk separately, then concatenate the summaries.

Answer

Increase the maxTokenCount parameter to allow the model to process more of the article.

Question 20

A healthcare company is building an application on Amazon Bedrock that uses a foundation model to answer patient questions about medications. The company must reduce the risk of harmful or inaccurate medical advice. Which TWO strategies should they implement? (Choose two.)

Accepted Answer

Use Amazon Bedrock Guardrails to filter harmful content and define denied topics related to specific medical advice.. Guardrails provide a managed safety layer that filters harmful content and blocks denied medical topics, while RAG grounds responses in approved clinical guidelines so answers reflect authoritative sources. Together they reduce both harmful outputs and factual inaccuracies, which is essential for a patient-facing medication assistant.

Answer

Disable all logging so patient interactions are not stored.

Answer

Fine-tune the model on a small set of patient questions and answers collected from previous chats.

Answer

Increase the model's temperature to make answers more diverse and comprehensive.