# Guide to SharePoint Data Access in Blockbrain
When integrating SharePoint with Blockbrain, users can choose between two main approaches:
* **A. Knowledge Management Indexing (EPL Pipeline)**
* **B. SharePoint Delegated Access via Agent**
This guide explains when to use each method, their pros and cons, and provides practical use-cases.
| Criteria | Knowledge Management Indexing (EPL Pipeline) | SharePoint Delegated Access (Agent) |
| -------------------- | -------------------------------------------- | ----------------------------------- |
| Data Volume | Small/Medium or Selective | Large |
| Cost | Higher (for large data) | Lower (selective access) |
| Answer Quality | High (Blockbrain controls search/indexing) | Limited (MS Graph API dependent) |
| Search/Analytics | Full | Limited |
| Use-Case | Comprehensive KM, Selective | Selective, cost-sensitive, agentic |
| Multi-Tool Workflows | No | Yes |
| Control of Search | Blockbrain | SharePoint (Microsoft Graph API) |
***
### Decision Criteria
#### **A. Knowledge Management Indexing (EPL Pipeline)**
**When to Use:**
* You need **high-quality, comprehensive search and answers** across your selected/specific SharePoint folders or files.
* Your SharePoint data volume is **manageable** (not extremely large).
* You want **specific documents/folders indexed** for advanced knowledge management features.
**How it Works:**
* The platform triggers an ETL pipeline to **index selected SharePoint data** (not the entire SharePoint).
* Data runs through Blockbrain’s **knowledge extraction and indexing pipeline**.
* Blockbrain controls the search and retrieval queries on its own index, enabling advanced and customizable search.
**Pros:**
* Best answer quality and search experience.
* Full content of indexed files is available.
* Enables advanced analytics and knowledge discovery.
* **Selective indexing:** You can choose only the folders/files you need.
**Cons:**
* **High cost** for large data volumes (gigabytes/terabytes).
* Indexing large SharePoint sites or many folders can cause **cost spikes**.
* Not suitable for massive datasets.
***
#### **B. SharePoint Delegated Access via Agent**
**When to Use:**
* Your SharePoint contains **large volumes of data** (many GBs or TBs).
* You want to **limit costs** by only connecting specific files or folders.
* You need **on-demand access** to selected SharePoint content.
* You want to combine SharePoint access with other agentic actions (e.g., drafting emails, cross-checking CRM, generating charts).
**How it Works:**
* The Agent connects to SharePoint with **delegated access**.
* You **choose which files/folders** to connect—no indexing or data extraction on our platform.
* **Extraction, indexing, and search are controlled by SharePoint (Microsoft Graph API), not Blockbrain.**
* Agentic tools can combine SharePoint actions with other workflows.
**Pros:**
* **Cost-effective** for large datasets.
* Flexible—connect only what you need.
* Avoids unnecessary indexing and storage costs.
* Can combine SharePoint access with other tools (e.g., email drafting, CRM checks, chart generation).
**Cons:**
* **Answer quality and search capabilities are limited by the Microsoft Graph API** (not by our platform).\
\&#xNAN;*Note: We cannot influence the search quality provided by the API. For most use cases, this is sufficient.*
* Not all content is available for AI-powered insights.
* Currently, **it does not support SharePoint search folders**.
More information on SharePoint access via Agent: [SharePoint Agent](https://app.gitbook.com/o/wESD8a1fR7H7ddwkYO61/s/IabFtGTeQzwfWCzp8vd6/~/edit/~/changes/451/for-admins/agents/sharepoint-agent).
***
### Use-Cases
#### **Use-Case 1: Small/Medium SharePoint Site**
* **Scenario:** A team wants to search and analyze project documents (few GBs).
* **Recommendation:** Use **Knowledge Management Indexing** for best results.
#### **Use-Case 2: Large Enterprise SharePoint**
* **Scenario:** An organization has several TBs of archived data but only needs to access current project files.
* **Recommendation:**
* Use **Selective Indexing** to index only the required folders via Knowledge Management, if advanced search is needed.
* Alternatively, use **SharePoint Delegated Access via Agent** for on-demand, cost-effective access.
#### **Use-Case 3: Cost-Sensitive Project**
* **Scenario:** A department has a tight budget and only needs to access a handful of files for a short-term project.
* **Recommendation:** Use **Delegated Access** to avoid unnecessary indexing costs.
#### **Use-Case 4: Cross-functional Tool Usage**
* **Scenario:** A user needs to read a file from SharePoint, draft an email with its content, and cross-check information in a CRM.
* **Recommendation:** Use **SharePoint Delegated Access via Agent** to enable multi-tool workflows.
***
### Recommendations
* **For best answer quality and full knowledge management:**
* Use **Knowledge Management Indexing** (EPL Pipeline) for small to medium data volumes, or selectively index only the folders you need.
* **For large datasets, cost control, or agentic workflows:**
* Use **SharePoint Delegated Access via Agent** to connect only what you need and combine with other tools.
**Tip:**\
Always assess your data volume, business needs, and workflow requirements before choosing the integration method.
