SharePoint + RAG: Build Your Own AI Assistant, Build a working AI assistant for SharePoint documents and lists using Python, SQL Server vector search, reranking & LLM.
What you’ll learn
-
Understand the fundamentals of Retrieval-Augmented Generation (RAG) and how it can improve access to information stored in SharePoint.
-
Build a RAG ingestion pipeline that retrieves content from SharePoint documents and lists and stores embeddings in SQL Server.
-
Use vector search in SQL Server to find SharePoint content relevant to a user’s question.
-
Improve retrieved results using a reranker before sending context to a Large Language Model (LLM).
-
Generate AI-powered answers based on information retrieved from SharePoint.
-
Run and use a complete RAG application with SharePoint, SQL Server, Python, and a simple web interface.
Course Description
What if you could ask questions about the information stored in SharePoint and get a direct, AI-generated answer instead of searching through documents, lists, and pages of search results?
In this short, hands-on course, you will learn how Retrieval-Augmented Generation (RAG) can be used with SharePoint Server to create an AI-powered question-answering solution.
We will start with the problem: traditional search is great for finding content, but sometimes what we really want is an answer. You will learn what RAG is, how it works, and how components such as embeddings, vector search, reranking, and Large Language Models (LLMs) work together.
Then we will build and run a complete RAG solution around the Microsoft technology stack.
Our SharePoint environment will contain both documents and list data. We will create an ingestion pipeline that reads this content, divides documents into smaller chunks, generates embeddings, and stores them as vectors in Microsoft SQL Server.
Next, we will build the query pipeline. A user’s question will be converted into an embedding and used to search SQL Server for relevant SharePoint content. We will then use a reranker to select the best results and provide that context to an LLM to generate the final answer.
Finally, we will put everything together in a simple web interface where you can ask questions and receive answers based on the information stored in SharePoint.
What you will learn
- Understand RAG and the problem it solves
- Understand embeddings and vector search without going deep into the mathematics
- Use SharePoint documents and lists as RAG data sources
- Store and search embeddings using Microsoft SQL Server
- Build a simple RAG ingestion pipeline with Python
- Retrieve and rerank relevant SharePoint content
- Generate answers using an LLM
- Run the complete solution through a simple user interface
- Understand the end-to-end architecture of a practical SharePoint RAG solution
This course is intentionally short and practical. You don’t need to become an AI researcher or study complicated machine-learning theory. The goal is to understand the main concepts, see how the pieces fit together, and finish the course with a working RAG project that you can explore and improve yourself.
Let’s give SharePoint a new way to answer questions!
Who this course is for:
- SharePoint administrators and developers who want to explore how RAG and generative AI can be integrated with SharePoint.
- Microsoft technology professionals interested in building practical AI solutions using SharePoint, SQL Server, Windows Server, and Python.
- Developers and IT professionals who want a hands-on introduction to RAG without going deep into AI or machine-learning theory.
- SharePoint professionals looking for a more conversational way to access information stored in documents and lists.
- AI and RAG beginners who want to understand the complete workflow; from ingestion and embeddings to vector search, reranking, and answer generation.
- Technical learners who prefer seeing a complete working project rather than studying RAG concepts only in theory.
