Databricks-GenAI-Assoc Design Applications Practice Question
A financial services company is designing a RAG assistant that must never return answers containing personally identifiable information from its knowledge base. The team plans to filter retrieved chunks before they are placed into the prompt. Which Databricks design element should they use to enforce this filtering consistently?
⚠ Common exam trap
The trap here is relying on prompt instructions or output filters for data protection, when the reliable control point is filtering retrieved chunks before prompt assembly.
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
✓
Store a sensitivity label or metadata field on each chunk in the Vector Search index and apply a metadata filter at query time to exclude restricted chunks.
Metadata filtering in Vector Search allows the application to exclude chunks based on attributes such as sensitivity level before they reach the prompt. This enforces the policy at the retrieval boundary, reducing both leakage risk and token usage. Prompt instructions and output regexes act later and cannot guarantee that sensitive content never enters the model context.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Add a system prompt instructing the model to ignore any personally identifiable information it sees in the context.
Why it's wrong here
A system prompt is a soft instruction and cannot guarantee that the model will omit sensitive content. The data would still be transmitted into the model context, creating exposure risk and potential leakage. Policy enforcement should happen before context assembly, not rely on model compliance.
- ✓
Store a sensitivity label or metadata field on each chunk in the Vector Search index and apply a metadata filter at query time to exclude restricted chunks.
Why this is correct
Vector Search supports metadata filtering, so tagging chunks with a sensitivity attribute and filtering at query time prevents restricted content from ever entering the prompt. This enforces the policy at retrieval, which is earlier and more reliable than post-generation redaction, and it scales as new documents are labeled.
- ✗
Run a post-processing regex on the model output to remove patterns that resemble personally identifiable information.
Why it's wrong here
Post-processing catches only what it can pattern-match and occurs after sensitive text has already been sent to the model and possibly logged. It also risks mangling legitimate answers. The requirement is to prevent PII from entering the prompt, which post-processing cannot guarantee.
- ✗
Grant the serving endpoint a service principal that lacks access to the sensitive tables so retrieval returns no restricted rows.
Why it's wrong here
Table-level access control does not apply to a prebuilt Vector Search index in the same way, and removing access could break legitimate retrieval of non-sensitive content. It also does not provide per-chunk filtering within an index that contains mixed sensitivity. Metadata filtering is the targeted mechanism for excluding specific chunks.
About these practice questions
Courseiva writes every Databricks-GenAI-Assoc question from scratch — 330 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. 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 Databricks exam blueprint
This Databricks-GenAI-Assoc practice question is part of Courseiva's free Databricks 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 Databricks-GenAI-Assoc exam.