AI-900 Practice Question: Describe Artificial Intelligence workloads and considerations

Answer analysis

Option-by-option breakdown

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

• ✗

  Scheduling when AI agents run to balance compute load across Azure regions

  Why it's wrong here

  Load-balancing compute across Azure regions is an infrastructure concern—it decides where and when to run workloads for latency or cost efficiency based on regional capacity. Multi-agent orchestration, by contrast, is a semantic runtime concern: it coordinates which autonomous AI agent performs which reasoning task, how information flows between them, and how their results are combined, irrespective of the underlying compute placement.

• ✓

  Coordinating multiple AI agents — planning tasks, delegating to specialists, and synthesising outputs

  Why this is correct

  Multi-agent orchestration is the runtime layer that plans a complex objective, decomposes it into subtasks, delegates each subtask to a specialized agent—such as a retrieval, tool-use, or code-generation agent—and then synthesizes their outputs into a coherent final answer. The orchestrator maintains shared state, handles inter-agent dependencies or conflicts, and can dynamically re-plan when an agent fails or returns unexpected results. This is the accepted meaning of 'orchestration' in AI agent systems.

• ✗

  Training a single model that can perform multiple specialised tasks simultaneously

  Why it's wrong here

  Training a single model to handle many specialized tasks simultaneously is typically multi-task learning or a unified model paradigm, where one neural network learns shared representations across objectives. This is a model-training approach, not agent orchestration—orchestration presumes multiple independently-defined agents (which may themselves be separately trained) cooperating at inference time, with the orchestrator managing their interaction rather than blending their training objectives.

• ✗

  Organising AI agent code in a Git repository for version control

  Why it's wrong here

  Organizing AI agent code with Git is a software engineering practice focused on version control, branch management, and change tracking. It is completely orthogonal to multi-agent orchestration, which is about the live execution and coordination of autonomous agents—planning, message passing, and task delegation—during runtime. Version control manages code history and collaboration among developers, not the behavioral coordination of AI agents.

Go deeper

Related to this question