Designing and Implementing Multi-Agent AI Solutions

The requirements ask for behavior that applies to every future conversation, improves routing for representative compound inputs, and standardizes final output with minimal application-code change. Few-shot instruction examples satisfy all three by demonstrating both the desired routing decision and the expected schema-compliant response for representative cases. Repository-wide constraints list rules but do not demonstrate how ambiguous compound requests should be handled. Per-request prompt cues are not durable across future conversations and would require application logic to inject them repeatedly. System role instructions define domains and objectives but provide less behavioral specificity than examples. Microsoft AI-500 prompt-engineering objectives explicitly include examples and dynamic prompt techniques for shaping complex agent behavior. Therefore C is the strongest design. The implementation should also preserve clear inputs and outputs around this step so that later agents receive only the information they require. This improves debuggability and keeps token, permission, and state growth under control as the workflow becomes more complex.

Official Microsoft reference: AI-500 Study Guide - advanced prompt engineering

The restriction belongs in the agent's MCP tool configuration because Microsoft Foundry supports an `allowed_tools` allowlist that controls which tools discovered from an MCP server are exposed to the agent. Configuring the allowlist to include only `docs_search` makes `deployment_delete` unavailable for model selection. This is stronger than adding a sentence to the agent instructions because instructions influence behavior but do not remove a dangerous tool from the callable surface. Project details describe resources rather than per-agent tool exposure, and a transient run setting is not the appropriate persistent configuration boundary for the registered MCP integration. Therefore C, the agent tool configuration, is the correct answer. Least privilege remains the governing principle: grant only the identity, data, tool, or deployment access required for the specific operation. The selected answer preserves that boundary while still allowing the workflow to satisfy its functional requirement. From a security and governance perspective, the control should be enforced at the narrowest platform boundary that can deterministically block or constrain the action. Relying only on prompt text is weaker because the model can still be induced to behave unexpectedly.

Official Microsoft reference: Microsoft Foundry agents - Model Context Protocol tools

Fraud-check, Policy-eligibility, and Document-summary are independent specialist activities, so running them concurrently reduces wall-clock time without changing their outputs. The Decision agent is different because the case study states that it depends on the results of those three agents. It therefore must execute only after the independent work completes. Microsoft Agent Framework's concurrent orchestration pattern is designed for independent participants whose results can be fanned in later, while dependency-driven work should remain ordered. Running all four agents concurrently would violate the Decision agent's data dependency. Running all four sequentially would preserve correctness but add unnecessary latency. Disabling memory would also change functionality unrelated to the performance bottleneck. The optimal graph is therefore a parallel fan-out of the three independent specialists followed by a fan-in and Decision step, which preserves behavior while minimizing end-to-end claim-approval duration. In production, add telemetry and regression tests around this behavior so changes to prompts, models, tools, or orchestration do not silently alter the intended contract. The selected approach is the one that best matches the platform's native execution semantics.

Official Microsoft reference: Microsoft Agent Framework - Concurrent orchestration

When multiple projects share one Foundry resource, grouping costs only by Azure resource cannot distinguish project consumption because the resource boundary is shared. Microsoft Foundry integrates with Azure Cost Management by attributing model usage to project metadata/tags, allowing a Cost Analysis view to filter shared resource spending by project. This supports showback or chargeback without requiring administrators to manually tag every usage event. Subscription-level service filters or meter grouping identify the service or billing meter but not which of the three business-unit projects generated the usage. Therefore a Foundry-resource-scoped Cost Analysis view filtered by the project tag is the configuration that both reconciles to Azure Cost Management and minimizes manual tagging. D remains correct. The evaluation should also preserve correlation identifiers and version information where possible so a failed score can be traced back to the exact agent, model, tool call, or retrieval step that produced it. This turns the metric into an actionable diagnostic rather than only a dashboard number.

Official Microsoft reference: Microsoft Foundry - Manage costs

Microsoft Foundry Agent Service is the managed option that best satisfies containerized production agent execution, dynamic scaling, and low administrative overhead. Hosted agents run in Foundry-managed container compute and scale according to workload demand while the platform manages much of the runtime, endpoint, identity, and operational plumbing. AKS could provide strong tenant/workload isolation and autoscaling, but it also requires Kubernetes cluster operations, policy, upgrades, node pools, and scaling configuration, which conflicts with the requirement to minimize administration as the tenant count grows. Azure Container Instances is simpler but lacks the same managed autoscaling model, and GPU virtual machines create the highest infrastructure burden. Therefore C is the strongest fit for a SaaS platform that wants production agent compute without owning the orchestration platform. From an architecture perspective, the selection also creates explicit ownership and boundaries that can be tested independently. That matters in multi-agent systems because implicit sharing or loosely defined authority often becomes the source of cross-agent coupling, security drift, and difficult incident diagnosis.

Official Microsoft reference: Microsoft Foundry Agent Service overview