Copilot Studio: The Ultimate Guide to Adding ALL Knowledge Sources (Files, SharePoint, Azure & More)

Welcome to the channel! In this comprehensive tutorial, we dive deep into the core of any powerful agent: Knowledge Sources.

If you want to build an intelligent agent in Microsoft Copilot Studio, you need to feed it the right information. In this video, we walk through the entire process of creating an agent from scratch and then systematically explore three (and more!) different options for connecting your data.

We start with the basics and move to advanced enterprise integrations, showing you exactly how to upload files, connect to cloud storage, and even set up Azure AI Search. By the end of this video, you'll know exactly how to make your Copilot smart and useful.

In this step-by-step guide, we cover:

Introduction: What are knowledge sources and why they are the backbone of your agent.

Creating a Basic Agent: A quick start in Copilot Studio.

Knowledge Source #1: Upload a File: The simplest method. We create a PDF document on computer hardware parts and upload it to test the agent.

Knowledge Source #2: OneDrive: How to connect your agent to files stored in the cloud, including browsing and using direct file URLs.

Knowledge Source #3: SharePoint: Connecting to your organization's document libraries by browsing items or pasting a file link.

Knowledge Source #4: Public Website: Pointing your agent to a public-facing website (like the official Power Automate docs) to use as a knowledge base.

Knowledge Source #5: Dataverse: Creating a custom table in Dataverse, importing data from a CSV file, and connecting it to your agent for structured Q&A.

Knowledge Source #6: Azure AI Search: A deep dive into setting up an enterprise-grade search index. We cover:

Creating a Knowledge Source and Blob Storage container.

Uploading files to Azure Blob Storage.

Setting up Text Vectorization with Microsoft Foundry (Azure AI Foundry) for advanced AI-powered search.

Creating a Knowledge Base and Index.

Connecting your Copilot Studio agent to your custom Azure AI Search index.

Final Testing & Configuration: We adjust the agent's instructions to control its behavior and ensure it only answers from the provided knowledge.

Why watch this video?
Whether you're a beginner or an experienced developer, this tutorial provides a clear, practical understanding of how to leverage all the major knowledge sources in Copilot Studio to build truly powerful and knowledgeable agents.

Timeline:
00:00 - Introduction to Copilot Studio and why knowledge sources are important
01:15 - Creating a basic agent and exploring the setup interface
02:30 - Overview of available knowledge source options
03:00 - Knowledge Source #1: Uploading a file (creating a PDF from DeepSeek)
05:15 - Testing the agent with the uploaded file
05:45 - Knowledge Source #2: Connecting OneDrive (browsing and using file URLs)
08:15 - Knowledge Source #3: Connecting SharePoint (browsing items and copying file links)
10:30 - Testing the agent with multiple knowledge sources
11:00 - Knowledge Source #4: Adding a public website (Power Automate official site)
12:00 - Knowledge Source #5: Setting up Dataverse (creating a table and importing CSV)
15:30 - Publishing and adding the Dataverse table to the agent
16:15 - Knowledge Source #6: Azure AI Search setup (creating storage container)
18:30 - Configuring text vectorization in Microsoft Foundry
21:00 - Deploying embedding models for Azure AI Search
22:45 - Creating knowledge source and knowledge base in Azure AI Search
24:30 - Adding Azure AI Search index to Copilot Studio
25:15 - Modifying agent instructions and final testing
27:00 - Summary and what's coming in the next video

Видео Copilot Studio: The Ultimate Guide to Adding ALL Knowledge Sources (Files, SharePoint, Azure & More) канала Code & Cloud