Medium AIF-C01 practice questions

A financial services company wants to build a chatbot that answers employee questions using their internal policy documents stored in Amazon S3. The team wants to minimize infrastructure management and avoid building a custom pipeline for chunking documents, generating embeddings, and storing them in a vector database. Which Amazon Bedrock capability best fits this need?

A company wants to give employees a conversational assistant that can answer questions using the company's internal documents stored across SharePoint, Confluence, and S3, while respecting existing user access permissions. The team wants a fully managed solution and prefers not to build or fine-tune their own model. Which AWS service best fits this requirement?

A contact center wants to help live support agents respond faster during customer calls. As the customer speaks, the company wants an AI tool that listens to the conversation, understands the caller's intent in real time, and automatically surfaces relevant knowledge-base articles and suggested responses to the agent. Which AWS generative-AI capability is purpose-built for this real-time agent-assist use case?

A retail analytics team wants business users to ask questions about sales data in plain English (for example, 'What were the top 5 products by revenue last quarter?') and receive charts and narrative summaries directly inside their existing business intelligence dashboards. Which AWS generative-AI capability is designed for this use case?

A financial services company is deploying a generative AI application on AWS and must provide its external auditors with formal documentation proving that the AWS services it uses meet SOC 2 and ISO 27001 compliance standards. Which AWS service should the team use to obtain these official third-party compliance reports and audit artifacts?

A media company uses a foundation model on Amazon Bedrock to generate SEO metadata descriptions for its archive of 2 million historical articles. The processing is a one-time bulk job with no user waiting for results, and the team wants to minimize cost. Which invocation approach best fits this workload?

A retail analytics team has 12 months of accumulated sales records stored in Amazon S3. Once every night, they want to score the entire dataset with a trained model to generate a report for the next business day. There is no requirement for immediate, per-request responses, and no application waits synchronously for individual predictions. Which inference approach best fits this need?

A travel company wants to build a generative-AI assistant that can answer customer questions and also complete multi-step actions such as looking up a booking in an internal database, checking flight availability via an existing API, and confirming a rebooking. The team wants a foundation model to reason about which steps to take and automatically call the right internal APIs to fulfill the request. Which Amazon Bedrock capability is designed for this use case?

A retail company is building a customer-facing chatbot on Amazon Bedrock. The chatbot must answer simple product questions with very low latency and at the lowest possible cost per request. The queries are short and do not require complex reasoning or long-form generation. The team is deciding which class of foundation model to select within Bedrock. Which selection approach best fits these requirements?

A financial services company wants to adapt a foundation model in Amazon Bedrock so it responds using the company's specific terminology and preferred answer style. They have a curated dataset of about 5,000 high-quality prompt-and-response pairs written by their domain experts. They want the model to learn to produce responses that match these labeled examples. Which Amazon Bedrock customization approach best fits this requirement?

A financial services company uses Amazon Bedrock foundation models for a customer-facing assistant. To satisfy an internal audit requirement, the compliance team needs to capture the full prompts submitted to the model and the complete model responses so they can be reviewed later for inappropriate content and retained in a company data store. Which approach BEST meets this requirement?

A marketing team at a retail company spends many hours each week manually writing product descriptions, social media captions, and promotional email drafts for thousands of catalog items. Leadership wants to understand the primary business value that adopting a generative AI solution would bring to this workflow. Which statement best describes that business value?

A financial services company uses Amazon Bedrock for a customer-facing assistant. To meet regulatory requirements, the compliance team must maintain a tamper-evident record of who invoked which foundation model, when, and from which account, so auditors can review this activity months later. Which AWS service should the team use to capture this audit trail of API activity?

A financial services company runs a fraud-detection model on a real-time SageMaker endpoint. The security team needs to be alerted automatically when the endpoint experiences an unusual spike in invocation errors or latency, which could indicate a denial-of-service attempt or a malfunctioning model. They want a native AWS solution that requires no custom polling code. Which approach best meets this requirement?

A media company wants to select between two candidate foundation models for a customer-facing product-description generator. They have assembled a fixed set of 500 product inputs with high-quality reference descriptions written by their editorial team. Before launch, they want an objective, repeatable way to compare which model produces output closest to their reference standard across the full test set. Which evaluation approach best meets this need?

A retail company deploys a machine learning model that classifies incoming customer support tickets into categories like 'Billing', 'Shipping', and 'Technical'. For each ticket, the model returns the predicted category along with a numeric value between 0 and 1 for each possible category. The support team wants to automatically route tickets only when the model is highly certain, and send low-certainty tickets to a human agent. Which characteristic of the model output should the team use to make this routing decision?

A legal firm wants to use a foundation model on Amazon Bedrock to summarize entire contract documents. Some contracts are extremely long, and the team notices that when they paste a very large contract into a single prompt, the model returns an error or ignores the later portions of the document. What is the most likely cause of this behavior?

A telecom company has already built and validated a machine learning model that predicts whether a customer is likely to cancel their subscription. Each night, a batch job runs the current customer records through this model to generate a churn-risk score for every account, which the retention team uses the next morning. Which phase of the ML process does this nightly batch job represent?

A European financial services company must comply with regulations requiring that all customer data and model inference processing remain within the EU. The compliance team asks how they should configure their Amazon Bedrock usage to meet this data residency requirement. What is the most appropriate action?

A data scientist at a logistics company is building a machine learning model to estimate package delivery times. Before training, she divides her historical dataset into three separate portions. She uses one portion to fit the model, a second portion to tune hyperparameters and compare model configurations during development, and a third portion that she holds back and uses only once, at the very end, to report an unbiased estimate of how the model will perform on new data. What is the third, held-back portion of data commonly called?

A team is building an image recognition system that automatically identifies defective parts on a manufacturing line. Their approach uses a model with many layers of interconnected nodes that automatically learn hierarchical features (edges, then shapes, then whole objects) directly from raw pixel data, without engineers manually specifying which features to extract. Which subfield of AI/ML does this approach best describe?

A retail company has a dataset of 50,000 past customer support emails, each already manually tagged by staff as either 'billing', 'shipping', or 'technical'. The company wants to build a model that automatically assigns one of these three categories to each new incoming email. Which type of machine learning approach best fits this task?

A media company wants to build a semantic search feature so that users can find articles related in meaning to a query, even when the exact keywords do not match. An AI engineer plans to convert each article and query into numerical vectors that capture semantic meaning, then compare them for similarity. Which generative AI concept is the engineer relying on?

A healthcare company sends patient data to a SageMaker real-time inference endpoint for diagnosis predictions. A compliance auditor requires that the patient data be protected while it travels over the network between the client application and the endpoint. Which measure directly addresses this requirement?

A healthcare company stores patient datasets in an Amazon S3 bucket that will be used to train an Amazon SageMaker model. Compliance auditors require that the training data be encrypted at rest using keys the company can manage, rotate, and audit. Which approach meets this requirement with the least operational overhead?