SharePoint Integration FAQ

This guide covers the most common questions regarding how Ebbot connects to your SharePoint environment, how security is handled, and what the experience looks like for your end-users.

---

Permissions & Connectivity

Does Ebbot get access to our entire SharePoint environment?

No. Ebbot only sees the specific sites you explicitly connect. You grant access to the Ebbot app one site at a time. We never have "Global" access to your SharePoint environment; you remain in total control of the data boundaries.

If I connect a "Hub Site," are sub-sites included automatically?

No. Giving Ebbot permission to read a Hub Site does not automatically grant it access to associated sub-sites. You must provide permissions and set up the connection for each specific site you want the bot to learn from.

Do we need to change our SharePoint layout or theme?

Not at all. Ebbot works entirely in the background via the SharePoint API. You don't need to move folders, change your navigation, or redesign your pages. However, keep in mind that the AI agent focuses on the text and data on a page rather than the visual design.

Best practice for sensitive data: If your site contains a mix of "public" knowledge and "highly sensitive" internal documents, we recommend creating a dedicated SharePoint site specifically for the AI agents knowledge. This creates a physical security boundary that is much easier to manage than filtering individual files.

---

How the scraper "thinks"

How is this different from our existing SharePoint Search (like PnP)?

Traditional search functions (like PnP) provide a list of links and hope the user finds the answer. Ebbot is a different beast: it actually reads the content within your documents and generates a direct, conversational answer.

• Search: "Here are 5 PDFs about travel policy."

• Ebbot: "You can claim up to $50 for dinner when traveling for business."

Why does the data sets sometimes contain the same document multiple times as a source?

Ebbot "reads" documents in small sections (chunks). If a user asks a question that is answered by three different paragraphs in the same manual, the AI agent will cite all three sections.

• Pro-tip: To keep your source list clean, try to keep your source documents modular and concise rather than uploading 100-page "Mega-Manuals."

---

User experience & targeting

Can users search specific sites (e.g., just the IT site)?

Yes, but this is managed via Datasets within Ebbot rather than the user selecting a site. You can group your SharePoint sites into different datasets. Depending on where the user is (e.g., on the IT Support page) or what their role is, we can tell the AI agent to only look for answers within a specific group of sites.

We have 30+ sites with different needs. Can we manage this?

Absolutely. Ebbot is designed to scale. We can configure the chat flow to be context-aware. For example, if a user is asking questions on the "Server Support" page, the AI agent can prioritize the Server site dataset. If they are on the general HR portal, it can look across the broader company dataset.

---

Troubleshooting & technical Notes

My SharePoint link is failing or says "No docs found."

If you are setting up a source and it isn't finding content, double-check that you are using the Site ID and not the SharePoint URL. The API requires the specific unique identifier for the site to pull data correctly.

Can I filter by SharePoint metadata?

The API connector currently reads the folder structure and file names rather than metadata tags. To restrict what the AI agent sees, we recommend using Path-based filtering (pointing the bot to a specific folder) or managing access via a dedicated Service Account.

Technical ID note: The App-ID used for the SharePoint connection is unique to your platform. Ensure your IT admin uses the specific ID provided in your Ebbot dashboard during the Azure authentication process.

---

Need more help? If you're unsure about your specific site architecture, feel free to reach out to [email protected]