Identify Benefits and Capabilities of Generative AI Solutions AB-731 Practice Questions (Page 7)

Questions 31–35

1. 31expert · hard

   A company has a generative AI system that summarizes legal documents. The system sometimes omits critical clauses. The team has already tried adding more instructions to the prompt, but the problem persists. They have a limited budget and need a solution that improves accuracy without retraining the model. What should they do?

   • AUse a different prompt template with more detailed instructions.
   • BSwitch to a larger language model with more parameters.
   • CUse a retrieval-augmented generation (RAG) approach that retrieves the full document and instructs the model to base the summary on the retrieved content.
   • DIncrease the model's temperature to encourage more thorough summaries.
   Reveal Solution Discussion Select an answer first
2. 32application · medium

   A healthcare organization is deploying a generative AI system that assists doctors with clinical documentation. The system will access patient records and generate summaries. The organization must ensure that only authorized medical staff can use the system and that patient data is protected. What should the organization prioritize?

   • AFocus on prompt engineering to improve the quality of the summaries.
   • BImplement strong authentication and role-based access control for the AI system.
   • CDeploy the system on-premises to avoid all security risks.
   • DUse a public cloud AI service without additional security configurations.
   Reveal Solution Discussion Select an answer first
4. 33application · medium

   A university wants to build an AI assistant that helps students find information from the university's course catalog, which is updated each semester. The assistant must always reflect the current catalog. Which approach best meets this requirement?

   • AUse a RAG solution that indexes the latest catalog and retrieves relevant sections for the model.
   • BUse a generic model and provide the catalog URL in the prompt.
   • CFine-tune the model on the current catalog each semester.
   • DStore the catalog in a static file and include it in the system prompt.
   Reveal Solution Discussion Select an answer first
5. 34expert · hard

   A company wants to deploy a generative AI assistant for its employees to query internal knowledge bases. The knowledge bases contain both structured data (e.g., databases) and unstructured data (e.g., documents). The assistant must provide accurate answers that combine information from both sources. What is the best approach?

   • AUse a simple keyword search over the documents and ignore the structured data.
   • BUse a separate AI system for each data type and combine the outputs manually.
   • CFine-tune the model on both structured and unstructured data.
   • DUse a RAG system that retrieves from both structured and unstructured sources and combines the results in the prompt.
   Reveal Solution Discussion Select an answer first
6. 35foundation · easy

   An organization wants to reduce the number of attempts needed to get a useful answer from a generative AI model. Which approach is most likely to improve the efficiency of the model's output generation?

   • AUsing a random seed for every generation
   • BIncreasing the temperature parameter to the maximum
   • CRemoving all examples from the prompt
   • DProviding a clear, detailed prompt with context and constraints
   Reveal Solution Discussion Select an answer first