Generative AI Leader Google Cloud's Generative AI Offerings Practice Question
A bank is deploying a retrieval-augmented generation application on Vertex AI so that a Gemini model answers employee policy questions using the bank's internal document repository. The team wants the model's responses to cite source documents and to reduce fabricated content. Which two capabilities should they implement to ground the model in the bank's own content? (Choose two.)
⚠ Common exam trap
The trap here is treating model retraining or safety filtering as a substitute for retrieval, when grounding in proprietary documents is fundamentally a retrieval-at-inference-time problem.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Vertex AI embeddings with vector search over the document corpus
Grounding responses in the bank's own content requires retrieving relevant internal passages and supplying them to the model. Indexing documents with Vertex AI Search or generating embeddings and retrieving them through vector search both provide that context, which improves factual accuracy and supports citations to source documents. Training or safety-filter adjustments do not retrieve proprietary content at inference time and therefore cannot satisfy the citation requirement.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Gemini safety filters configured to block sensitive topics
Why it's wrong here
Safety filters restrict categories of harmful output and are important for responsible deployment, but they do not connect the model to the bank's document repository or produce citations. Configuring them changes what the model will refuse to generate rather than what knowledge it draws upon, so it does not accomplish the grounding and attribution goals described in the scenario.
- ✓
Vertex AI embeddings with vector search over the document corpus
Why this is correct
Generating embeddings for the bank's documents and storing them in a vector index such as Vertex AI Vector Search allows retrieval of semantically similar passages for each query. Those retrieved passages become the grounding context for the Gemini model, enabling responses that reflect internal policy content and can reference the retrieved sources, which fulfills the grounding and citation objectives.
- ✗
Grounding with Google Search
Why it's wrong here
Grounding with Google Search connects model responses to public web results, which is useful for current events and general knowledge. It does not expose the bank's internal document repository, so responses could not cite internal policy documents. Because the requirement is to ground answers in proprietary content, public web grounding fails to satisfy the scenario even though it is a legitimate grounding technique.
- ✓
Vertex AI Search grounding with the internal document index
Why this is correct
Vertex AI Search can index the bank's internal documents and return relevant passages that are supplied to the Gemini model as grounding context, which both improves factual accuracy and enables citations back to source documents. This directly addresses the requirement to answer from the bank's own repository while reducing hallucination and supporting attribution to the originating content.
- ✗
Vertex AI Pipelines scheduled retraining of the Gemini model
Why it's wrong here
Scheduled retraining or tuning changes model weights but does not by itself retrieve and cite the bank's documents at inference time, and it is far more costly to keep current as policies change. The scenario asks for grounded responses with source attribution, which is a retrieval concern rather than a training concern, so retraining does not deliver the required citations or reduce fabrication as directly.
Go deeper
Related to this question
About these practice questions
One of 1,008 original Generative AI Leader practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Google Cloud exam blueprint
This Generative AI Leader practice question is part of Courseiva's free Google Cloud certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the Generative AI Leader exam.