Question 19
Domain 3: Application DevelopmentWhat is the main benefit of asking an LLM for structured output with a fixed schema in an application pipeline?
Correct answer: B
Explanation
A fixed schema forces the LLM to return data in a predictable format, which makes it easier for later steps to parse without ambiguity. It also supports validation because the application can check the output against the expected fields and types before using it.
Why each option is right or wrong
A. It guarantees zero hallucinations
Schemas improve format consistency, but models can still generate incorrect or fabricated content.
B. It makes downstream parsing and validation more reliable
A fixed schema constrains the model’s response to predefined fields and types, so the application can parse the output deterministically instead of relying on free-form text. In a pipeline, that predictability reduces ambiguity and lets validation logic check required keys, data types, and missing values before the result is passed to later steps.
C. It removes the need for prompts
Prompts still guide the model; schema constraints do not replace instruction design.
D. It always reduces token cost
Structured output may help efficiency, but token usage does not always decrease.