Home / GenAI Engineer practice test / Application Development / Question

Databricks Generative AI Engineer Associate · free practice question

Consider a scenario where an agent within a multi-agent system needs to access structured business information stored in a Delta table, solely through natural…

Exam-style question on Application Development from the free GenAI Engineer practice test, with the answer and a full explanation below.

Updated November 11, 2026

Application Development

Consider a scenario where an agent within a multi-agent system needs to access structured business information stored in a Delta table, solely through natural language queries, bypassing direct SQL code generation. The organization already has a Databricks AI/BI Genie Space pre-configured for this specific Delta table. What is the MOST effective method for integrating this agent to facilitate data retrieval using natural language?

  1. AConnect the agent to the Delta table through a Spark SQL tool so that it writes and runs its own SQL queries against the table.
  2. BDeploy a separate LLM fine-tuned on the table's schema to generate SQL, and expose it to the agent as a callable microservice.
  3. CAdd the existing Genie Space as an agent tool through its conversational API, so the agent sends questions and gets structured results.
  4. DBuild a Vector Search index over the Delta table and have the agent retrieve matching rows as text chunks by semantic similarity.
Answer & explanation

Correct answer: C

WHY C is correct: Databricks AI/BI Genie Spaces are purpose-built for natural-language querying over structured Delta tables. The Genie conversational API allows agents to send natural-language questions and receive verified, structured results — it handles SQL generation, execution, and result formatting internally. Integrating an existing Genie Space as an agent tool is the recommended pattern for enabling multi-agent structured data retrieval.

WHY NOT A: Having the agent generate raw SQL autonomously bypasses the already-configured Genie Space, replicates its functionality with additional SQL generation reliability risks, and requires direct Spark execution access from the agent's environment.

WHY NOT B: Building and hosting a separate SQL-generation microservice adds significant complexity and maintenance overhead, and is redundant when a Genie Space already provides this capability with Databricks-native governance and integration.

WHY NOT D: Vector Search is optimized for semantic retrieval over unstructured text. Returning Delta table rows as text chunks loses the structured, queryable nature of the data and cannot support filtered queries, joins, or aggregations.

Take the full GenAI Engineer practice test free →

Would you pass today?

This question is one of 45 in the free, timed Databricks Generative AI Engineer Associate practice test. Take it and get a section-by-section score.

Start the free test →