AB-731 practice question 16 of 35

Contoso Health is building a “Clinical Insights Copilot” for its operations and analytics teams. Requirements • R1: Produce a structured…

Full set: 389 questions
Contoso Health is building a “Clinical Insights Copilot” for its operations and analytics teams. Requirements • R1: Produce a structured, citation-rich briefing that blends clinical guidelines, internal policy docs, and recent research articles from the web on a given topic. • R2: For a set of attached CSV and Excel files (admissions, readmissions, length of stay), find trends, outliers, and key drivers across regions and time. • R3: Keep the user experience simple for clinicians: they should choose the right agent based on the task rather than configuring tools manually. Proposed solution • Use the Researcher agent when clinicians need a topic-level briefing that synthesizes web and work content into a report they can share. • Use the Analyst agent when analysts need to upload admissions and readmissions datasets and ask questions like “What changed vs. last quarter?” and “Where are outliers?” • Provide simple usage guidance so clinicians know “Researcher for narrative research; Analyst for dataset deep dives.” Does this mapping of requirements to Researcher and Analyst align with how these agents are intended to be used?
Answer options
A: Yes Correct answer
B: No
Correct Answer:
A. Yes
Explanation

Yes, the mapping is correct. The proposed solution effectively uses two distinct and appropriate AI agent archetypes to meet the specified requirements. The "Researcher" agent pattern is ideal for Requirement R1, which involves synthesizing information from diverse, often unstructured sources (documents, web) into a narrative briefing. The "Analyst" agent pattern is perfectly suited for Requirement R2, which focuses on interpreting structured data from files (CSV, Excel) to identify trends and anomalies. This clear separation of concerns simplifies the user experience (R3) by providing task-specific agents.

Why Incorrect

B. The mapping is logical and aligns with standard AI agent design patterns, where different agents are specialized for different types of tasks (e.g., knowledge synthesis vs. data analysis).

References

1. Microsoft Learn. "Microsoft Copilot Studio - Overview." learn.microsoft.com. The documentation on Copilot Studio explains how to build specialized copilots (agents) for specific tasks or audiences, supporting the concept of creating distinct "Researcher" and "Analyst" agents for different functions.

2. Wang, L., et al. (2023). "A Survey on Large Language Model based Autonomous Agents." arXiv preprint arXiv:2308.11432. Section 3, "Constitution of LLM-based Agent," discusses the modular design of agents, including planning and tool use, which supports specializing agents for tasks like research (information retrieval tools) versus analysis (code execution/data analysis tools). https://doi.org/10.48550/arXiv.2308.11432

Scroll to Top