Generative AI Leader Intermediate Practice Exam: Medium Difficulty 2025

A support team wants an LLM-powered assistant that answers questions using the company’s internal policy documents. They notice the model sometimes invents policy details. They want to reduce hallucinations without retraining a model. What is the most appropriate approach on Google Cloud?

A marketing team is generating product descriptions. They need consistent brand voice and must avoid disallowed claims. Which prompt strategy best meets these requirements?

A product manager wants to evaluate two prompt templates for a customer-email summarization workflow. They need a repeatable way to compare quality and reduce subjectivity. What is the best approach?

A retailer wants to build a GenAI chat experience in their mobile app. Requirements: integrate with Google’s foundation models, apply enterprise security controls, and connect to internal product and order data stored on Google Cloud. Which solution is most appropriate?

A team is prototyping a document Q&A assistant. They have thousands of PDFs in Cloud Storage and want to search semantically across documents and feed relevant passages to an LLM. What Google Cloud capability best supports this pattern?

A company is deploying a GenAI API for multiple internal teams. They want to enforce consistent safety settings, control access by team, and monitor usage. Which approach best aligns with Google Cloud best practices?

A legal team wants to summarize contracts and extract key clauses. Accuracy is critical, and they need traceability to the source text to support review. What should the solution include?

A customer service organization wants to reduce average handle time by suggesting draft responses to agents during chats. They want to mitigate risk while still improving productivity. What is the best deployment approach?

A bank is using GenAI to help summarize customer interactions. They must prevent exposure of sensitive personal data and comply with internal policy. Which combination of controls is most appropriate?

A company launches a GenAI feature and later discovers it produces biased recommendations for certain user groups. They want an operational plan to detect and reduce this risk over time. What should they do?