Question 40 — AB-620 Practice Exam 4
For each Copilot Studio knowledge requirement, drag it to the appropriate box in the Correct Match column to match it with the most suitable knowledge source type.
Each item may be used once, more than once, or not at all.
| Option | Correct Match |
|---|---|
| Ground answers in documents and pages stored in SharePoint libraries | [Gap 1] |
| Search structured records and tables maintained in a Power Platform data environment | [Gap 2] |
| Use documents added directly to the Copilot Studio agent | [Gap 3] |
| Ground responses in publicly accessible web content | [Gap 4] |
| Query vector indexes for semantic or hybrid retrieval scenarios | [Gap 5] |
| Use vector-indexed OneDrive or SharePoint files in Dataverse together with third-party knowledge-base articles available through supported connectors | [Gap 6] |
- Computer use
- Prebuilt connectors
- SharePoint
- Dataverse
- Public websites
- Unstructured data
- Azure AI Search
- File uploads
Correct answers
- Gap 1: SharePoint
- Gap 2: Dataverse
- Gap 3: File uploads
- Gap 4: Public websites
- Gap 5: Azure AI Search
- Gap 6: Unstructured data
Explanations
Correct Matches
SharePoint – SharePoint knowledge sources are designed for documents and pages stored in SharePoint libraries.
Dataverse – Dataverse knowledge sources provide access to structured tables and records maintained in a Dataverse environment.
File uploads – File-upload knowledge sources use documents that are added directly to the agent.
Public websites – Public website knowledge sources allow the agent to ground answers in content that is publicly accessible on the web.
Azure AI Search – Azure AI Search is suitable when semantic or hybrid retrieval must operate against vector indexes.
Unstructured data – Unstructured data can include OneDrive or SharePoint files synchronized and vector-indexed in Dataverse, together with third-party knowledge-base articles made available through Power Platform connectors.
Unused Item(s)
Prebuilt connectors – Prebuilt connectors are action-oriented tools for connecting to supported services; they do not represent any of the knowledge-source patterns described in the table.
Computer use – Computer use interacts with graphical interfaces when an API is unavailable; it is a tool mechanism rather than a knowledge source.