Integrating Microsoft SharePoint as a Data Source for Amazon Bedrock Knowledge Bases

Introduction

In today’s data-driven world, organizations need powerful tools to generate contextually relevant and customized responses. Amazon Bedrock Knowledge Bases provide just that by offering foundation models (FMs) and agents capable of leveraging Retrieval Augmented Generation (RAG) from a company’s private data sources. This fully managed RAG solution enhances the accuracy and relevance of generated responses.

Microsoft SharePoint as a data source into Amazon Bedrock Knowledge Bases opens new possibilities for enhancing AI-driven insights and decision-making. SharePoint is widely used for storing, organizing, and sharing critical internal data, making it a valuable resource for feeding foundational models with the right context. By connecting SharePoint with Amazon Bedrock, organizations can seamlessly access and retrieve structured and unstructured data, enabling the Bedrock models to generate highly relevant, contextual responses tailored to the specific needs of the business.

Solution Overview

The SharePoint data source within Amazon Bedrock Knowledge Bases offers several key features that enhance the integration of SharePoint content:

• It enables direct access to the data stored in SharePoint. With an RAG architecture, it queries SharePoint to retrieve the most relevant information, providing contextualized responses based on user input.
• It supports extracting structured data, metadata, and other document information from SharePoint, delivering relevant search results in response to user queries.
• It allows for ongoing synchronization of incremental content updates from SharePoint, ensuring that the knowledge base remains current.
• It offers source attribution for the responses generated by the foundation models, ensuring transparency and traceability of the data source.

Implementation Steps

Create a knowledge base and connect to the data source:

1. In the Amazon Bedrock console, navigate to the side panel and select Knowledge Bases.
2. Choose Create knowledge base.

3. In the Knowledge Base details section, you can modify the default name and add a description for your knowledge base.

4. Under the AWS IAM Permissions section, select an AWS IAM role that grants Amazon Bedrock access to other AWS services. You can allow Amazon Bedrock to automatically create the service role or choose a custom role you’ve already set up.

5. In the Choose Data Source section, select SharePoint as your data source.

6. Optionally, you can add tags to your knowledge base.

7. Click Next to proceed.

8. In the Name and Description section, you can optionally change the default name of the data source and provide a description.

9. In the Source section, enter the following details:

• For Site URLs, input the URLs of the sites you want to crawl and index for RAG purposes.
• For Domain, specify the domain name associated with the data source.
• In the Advanced Settings, leave the default selections as they are.

When converting your data into embeddings, Amazon Bedrock encrypts the data using an AWS-owned and managed key by default. If you prefer to use your AWS Key Management Service (AWS KMS) key, select Customize Encryption settings (Advanced) and choose your key.

You can also choose a data deletion policy for your data source from the following options:

• Delete – Removes all underlying data associated with the data source from the vector store when the knowledge base or data source resource is deleted. The vector store itself is not deleted, only the data. This option is disregarded if an AWS account is deleted.
• Retain – Keeps all underlying data in the vector store when the knowledge base or data source resource is deleted.

10. The Authentication section sets the supported authentication method to OAuth 2.0.

• For Tenant ID, enter your tenant ID.
• For the AWS Secrets Manager Secret, input the secret from AWS Secrets Manager.

11. In the Metadata Settings section, select any content types you want to include or exclude.

12. In the Content chunking and parsing section, select Default.

13. Choose Next.

14. In the Embeddings model section, select Titan Embeddings G1 – Text or another embedding model as appropriate.

15. In the Vector database section, select Quick Create a new vector store to create a vector store for the embeddings.

16. Choose Next.

17. On the Review and Create page, verify your selections and choose Create.

18. To initiate this process, select your data source on the knowledge base details page and choose Sync.

Conclusion

By enhancing accessibility to knowledge, this integration can significantly boost productivity, improve decision-making, and foster better knowledge sharing within the organization.

Drop a query if you have any questions regarding Amazon Bedrock and we will get back to you quickly.

About CloudThat