Databricks Generative AI Engineer Associate Practice Question 1: Agent Bricks Selection
Databricks Certified Generative AI Engineer Associate · Test 1 · question 1 of 10
An HR team keeps several hundred policy PDFs in a Unity Catalog volume. They want a chatbot that employees can ask questions such as "How many days of parental leave do I get?", and every answer must cite the policy document it came from. The team has little coding capacity and wants to use a Databricks Agent Bricks option rather than write a custom agent.
Which Agent Bricks option best fits this requirement?
Show answer and explanation
Answer:
Knowledge Assistant is the Agent Bricks option for building a question-and-answer chatbot over an organization's documents. It returns answers with citations and accepts Unity Catalog volumes, tables, or AI Search indexes as sources. That matches every requirement in the stem: free-form questions, cited answers, and a low-code build.
Option A is wrong. Information Extraction turns unstructured documents into structured JSON that follows a schema you define (it is built on ai_extract). That suits pulling fields out of invoices or contracts, not open-ended conversational Q&A.
Option B is wrong. A Supervisor Agent orchestrates *other* agents and tools, such as Genie agents, agent endpoints, Unity Catalog functions, and MCP servers. It would still need a document Q&A agent underneath it, so on its own it adds a layer without solving the problem.
Option D is wrong. Genie spaces answer natural-language questions about *structured* tabular data by generating SQL. They are not a retrieval chatbot over a folder of PDFs.
Reference: Use Knowledge Assistant to create a high-quality chatbot over your documents