Skip to main content
  1. Public practice questions
  2. AB-620 Practice Exams | AI Agent Builder Associate
  3. AB-620 Practice Exam 1

Question 44 — AB-620 Practice Exam 1

Fill in the blanks by selecting the correct options from the dropdown menus.

When authoritative content already resides in SharePoint, directly uploaded files, or Dataverse, begin with [Gap 1]. Non-Microsoft enterprise content that must support semantic indexing and citable responses should use a [Gap 2]. Data that cannot be replicated and must reflect its current state should use a [Gap 3]. When an organization maintains its own vector index, the appropriate knowledge integration is [Gap 4].

  1. Power Platform connector for real-time knowledge
  2. Agent flow
  3. built-in knowledge sources
  4. Azure AI Search
  5. REST API tool
  6. Copilot connector

Correct answers

  1. Gap 1: built-in knowledge sources
  2. Gap 2: Copilot connector
  3. Gap 3: Power Platform connector for real-time knowledge
  4. Gap 4: Azure AI Search

Explanations

Correct:

Built-in knowledge sources: SharePoint, file uploads, and Dataverse provide direct knowledge-source choices when the content already resides in those supported systems.

Copilot connector: Copilot connectors are suited to non-Microsoft enterprise content that needs semantic indexing and support for citable responses.

Power Platform connector for real-time knowledge: A real-time knowledge connector queries current source data without requiring the data to be replicated, making it appropriate for freshness-critical requirements.

Azure AI Search: Azure AI Search is appropriate when the organization manages the vector index used for retrieval.

Wrong:

REST API tool: A REST API tool is an action integration for directly exposing API operations to Copilot Studio. It is not the knowledge pattern represented by these grounding requirements.

Agent flow: An agent flow coordinates deterministic sequences of actions rather than providing the knowledge-retrieval patterns required by these scenarios.