AI-900 Practice Question: Describe features of generative AI workloads on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Technical error messages returned by the API when something goes wrong

  Why it's wrong here

  This option mistakes API error handling for the system message feature. When a request fails, Azure OpenAI returns an HTTP status code and a structured error body that describe authentication, rate-limit, or invalid-parameter problems. A system message, however, is not a reactive response to a failure; it is a proactive instruction included in the chat completions request that tells the model how to behave. Error messages are generated by the service after a request, whereas a system message is supplied by the developer before the model generates output.

• ✓

  Developer-provided instructions that define the AI's role and behavioral constraints for a session

  Why this is correct

  This is correct because a system message is developer-supplied metadata at the start of a chat completions request that defines the assistant's persona, scope, style, and session-wide behavioral rules. In Azure OpenAI, this message carries higher-level directives that influence how the model interprets and responds to every subsequent user turn, such as 'You are a helpful AI that answers only about Azure services.' It is not an error, alert, or notification; it is the central mechanism for configuring model behavior for a given conversation session.

• ✗

  Messages sent by the operating system to alert of resource usage

  Why it's wrong here

  This option confuses OS-level infrastructure monitoring with the LLM API's system message. Messages about CPU, memory, or disk usage are generated by the operating system or monitoring agents and have no connection to the natural-language instructions sent to Azure OpenAI. A system message is part of the conversational AI configuration that instructs the model on how to behave within a session, not an alert emitted by the underlying machine. These belong to fundamentally different layers: one is host telemetry, the other is prompt-level guidance.

• ✗

  Notifications sent to users when the AI service is experiencing issues

  Why it's wrong here

  This option confuses external operational alerts with the system message in an Azure OpenAI API request. Service health and status notifications are delivered through Azure Service Health / status pages to administrators, not as a parameter inside the conversation payload. In contrast, a system message is a model-level behavioral configuration sent in the API request to set the assistant's overall tone and constraints for the session. These are two completely separate mechanisms: one monitors Azure infrastructure, the other shapes model responses.

Go deeper

Related to this question