Question 47 — AB-620 Practice Exam 2
An organization maintains a specialized regulatory document collection in its own vector index outside Microsoft Graph. The organization wants Copilot Studio to use the index directly while retaining control over the indexing pipeline, embedding model, and relevance tuning. Which knowledge-source approach meets these requirements? (Select one answer)
Copilot connector
Microsoft Search
Real-time Power Platform connector
Azure AI Search
Correct answers
Azure AI Search
Explanations
Correct:
Azure AI Search
Azure AI Search allows Copilot Studio to connect to an organization-managed vector index outside Microsoft Graph. It is appropriate for semantically rich, unstructured collections that require custom control over indexing, embeddings, and relevance tuning.
Wrong:
Copilot connector
A Copilot connector ingests external content into an index managed by Microsoft Graph. It does not satisfy the requirement to retain the organization's existing custom vector index outside Microsoft Graph.
Microsoft Search
Microsoft Search can surface content that has been made searchable through Microsoft Graph, including content indexed by Copilot connectors. It does not provide the required direct integration with an organization-controlled Azure AI Search vector index.
Real-time Power Platform connector
A real-time Power Platform connector retrieves current data through on-demand API calls without creating a separate index. This pattern is suited to live records rather than a specialized document collection that already uses a custom semantic vector index.