Question 43 — AB-620 Practice Exam 5
Fill in the blanks by selecting the correct options from the dropdown menus.
A supported external service already has an appropriate Power Platform connection, so the preferred action pattern is a [Gap 1]. A proprietary REST API must be reusable by several agents and no centrally managed tool server is planned, so the organization should invest in a [Gap 2]. A proprietary API is needed only for a one-off Copilot Studio integration where full connector packaging is unnecessary, so a [Gap 3] is appropriate. A process requires several actions to execute in a defined sequence, so it should use an [Gap 4]. An enterprise plans to manage a common set of tools through a central server consumed by multiple agents, so it should use [Gap 5].
- agent flow
- prebuilt connector
- MCP
- REST API tool
- Copilot connector
- Azure AI Search
- custom connector
Correct answers
- Gap 1: prebuilt connector
- Gap 2: custom connector
- Gap 3: REST API tool
- Gap 4: agent flow
- Gap 5: MCP
Explanations
Correct:
Prebuilt connector: When an appropriate prebuilt connector already exists, it provides the direct integration path without requiring a custom API wrapper.
Custom connector: A custom connector is appropriate when a proprietary API needs reusable integration across multiple agents and centralized MCP server management is not the requirement.
REST API tool: A REST API tool suits a one-off proprietary API integration where connector packaging is not justified. REST API tools are scoped to Copilot Studio rather than acting as shared Power Platform resources.
Agent flow: Agent flows provide deterministic orchestration when multiple actions must execute according to a defined sequence.
MCP: Model Context Protocol is appropriate when an organization has or plans a centrally managed tool server that multiple agents can consume. For isolated or exploratory integrations without such infrastructure, a connector or REST API tool can be faster to establish.
Wrong:
Copilot connector: A Copilot connector is a knowledge integration pattern for enterprise content that needs semantic indexing and citable retrieval rather than an action pattern for these requirements.
Azure AI Search: Azure AI Search supports knowledge retrieval against organization-managed vector indexes rather than the action integration requirements represented by these scenarios.