Implement generative AI and agentic solutions 2.1 Build generative applications by using Foundry

A Foundry IQ knowledge base uses a Remote SharePoint knowledge source (preview).

Study guide skill: Implement retrieval-augmented generation (RAG) in an application

Question B6-014

A Foundry IQ knowledge base uses a Remote SharePoint knowledge source (preview). Users now require the references in each response to include service-generated citation URLs (preview) that resolve to the documents in the index, and the SharePoint content may be ingested into Azure AI Search. What should you do?

Select one answer.

  • A Replace it with an Indexed SharePoint knowledge source (preview), which generates an indexer pipeline and a backing search index ✓ Correct answer
  • B Raise the retrieval reasoning effort (preview) to medium
  • C Add a Web knowledge source to the knowledge base
  • D Set alwaysQuery to true on the Remote SharePoint knowledge source

Answer: A

Citation URLs (preview) come from an indexed knowledge source's backing search index; remote knowledge sources have no backing index, so they don't return them. Indexed SharePoint (preview) generates an indexer pipeline from a SharePoint site, so its results can carry citation URLs. alwaysQuery only includes a source in every query, a higher reasoning effort adds an iterative pass, and Web is another remote source.

Deep dive

How it works

Foundry IQ knowledge sources come in two families:

  • Indexed sources have a backing search index in Azure AI Search. Because of that index, references in a response can include service-generated citation URLs (preview) that resolve to their documents in the index.
  • Remote sources query content where it lives. They have no backing index, so they don't return citation URLs.

Remote SharePoint (preview) retrieves content from SharePoint at query time, which is why it can't meet the new requirement. Indexed SharePoint (preview) generates an indexer pipeline from a SharePoint site, which is allowed here because the content may be ingested into Azure AI Search. Switching to it gives the knowledge base a backing index, and with it, citation URLs.

The other options, and when you'd use them

  • Raise retrieval reasoning effort: The reasoning effort controls how much LLM processing is applied to each query. It doesn't give a remote source a backing index.
  • Add a Web knowledge source: Web is a remote source that retrieves real-time grounding data from Microsoft Bing. It has no backing index for your SharePoint documents.
  • alwaysQuery = true: It includes a source in every query, regardless of reasoning effort. The source stays remote, so still no citation URLs.

Exam tip

Citation URLs need a backing index, so you need an indexed knowledge source. Remote (SharePoint remote, Web) = no citation URLs.

References

Last reviewed:

Ready to test yourself? The quiz asks this and every other AI-103 question, with instant feedback. Practice this in a quiz →

All Generative AI & agents questions · AI-103 study guide

Thanks for your feedback!