Building a copilot: Azure AI Studio or Copilot Studio
summary
Builds two copilots for Contoso Outdoor: a customer assistant on Azure AI Studio using GPT‑4/GPT‑4o, CLIP multimodal embeddings stored in Azure AI Search vectors, and Prompt Flow orchestration to retrieve, augment, and respond with grounded product/order info; and an employee assistant in Copilot Studio using Power Platform connectors and a custom API connector (get orders, get order, update order) published to Teams with generative actions. A cross‑product flow posts to Prompt Flow to generate shipment emails; the integration works conceptually but HTTP/email steps show intermittent errors during the demo.
AI-written summary · about
chapters (10)
- 0:00Introduction to AI Tools
- 2:18Understanding Co-pilot Studio
- 4:22Exploring Azure AI Studio
- 5:48Customer Assistant Demo
- 12:13Implementing RAG Pattern
- 15:51Prompt Flow and Safety
- 21:29Employee Assistant Scenario
- 23:55Co-pilot Studio Actions
- 31:51Dynamic Actions in Teams
- 38:55Combining AI Studio and Co-pilot
related talks
46mBuild and maintain your company Copilot with Azure ML and GPT-4
01h 58mBuild, Evaluate and Deploy a RAG-Based Retail Copilot with Azure AI
49mBuild your custom Copilot with Microsoft Copilot Studio
45mThe era of the AI Copilot
15mAI Studio - Creating and scaling your custom copilots
27mBuild a RAG-based Retail Copilot Code-First on Azure AI Foundry
46mAI innovation in the Microsoft Power Platform
56mBuild your own CoPilot