AI-103 Certification Practice Question #22

Question

You have a Microsoft Foundry project that contains a customer support agent. The agent calls an internal knowledge API tool before generating responses.
Users report the following issues:
Some requests take more than 15 seconds to complete.
Some responses are incorrect, even when the knowledge API returns the expected data.
You need to inspect individual agent runs to view the ordered sequence of large language model (LLM) calls, tool invocations, and timing information.
Which observability capability should you use?

Accepted Answer

The requirement is to inspect an individual agent run and see the ordered sequence of LLM calls, tool invocations, and timing — this is the textbook definition of tracing in Microsoft Foundry observability. MS Learn states tracing captures the execution flow with spans for each LLM call and tool invocation, including timing, so you can pinpoint slow steps (the 15-second calls) and see why a response was wrong even when the tool returned good data (e.g., the model ignored the tool output). Monitoring (B) is the closest distractor but operates at the aggregate/dashboard level across many runs rather than letting you step through one run's ordered spans. Confidence is high because the Suggested Answer aligns precisely with the documented capability; the only deduction is the absence of community confirmation.

More AI-103 practice questions