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

Question 40 — AB-620 Practice Exam 1

A Copilot Studio agent requires two knowledge integrations. The first must search non-Microsoft enterprise content such as an IT service management repository after that content is indexed into Microsoft Graph, and responses should support semantic retrieval and citations. The second must query operational data directly from its source at runtime so that current values can be returned without replicating or moving the data. Which two knowledge source types meet these requirements? (Select two answers)

  1. Copilot connectors

  2. Power Platform connectors (real-time, preview)

  3. Azure AI Search

  4. File uploads

Correct answers

  1. Copilot connectors

  2. Power Platform connectors (real-time, preview)

Explanations

Correct:

Copilot connectors

Copilot connectors bring supported non-Microsoft enterprise content into Microsoft Graph indexing and support semantic search and cited responses.

Power Platform connectors (real-time, preview)

Real-time Power Platform connector knowledge sources query live source data at runtime without requiring the data to be replicated or moved.

Wrong:

Azure AI Search

Azure AI Search is designed for vector indexes and semantic or hybrid search scenarios. It does not satisfy the requirement to use Microsoft Graph-indexed enterprise content together with a separate live, no-data-movement connector pattern.

File uploads

File uploads make documents added directly to an agent available as knowledge. They do not provide runtime access to live operational data or Microsoft Graph indexing of an external enterprise repository.