Merge pull request #51049 from MicrosoftDocs/NEW-forging-voices-from-data

New forging voices from data -> main -- ASAP

Author: Albertyang0

learn-pr/philanthropies/forging-voices-from-data/1-introduction.yml

@@ -0,0 +1,15 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.introduction
+title: Introduction
+metadata:
+  title: Introduction
+  description: "In this unit, learners explore the full lifecycle of designing, customizing, and deploying AI-powered dialogue systems using Azure OpenAI Studio. They learn to select models, fine-tune them with custom data, personalize tone, and refine outputs through real-time testing. Keywords: Azure OpenAI Studio, dialogue systems, GPT-4, custom AI assistants, AI fine-tuning."
+
+  ms.date: 4/28/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+  
+durationInMinutes: 5
+content: |
+  [!include[](includes/1-introduction.md)]

learn-pr/philanthropies/forging-voices-from-data/2-azure-openai-studio-for-dialogue-systems.yml

@@ -0,0 +1,15 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.azure-openai-studio-for-dialogue-systems
+title: Azure OpenAI Studio for dialogue systems
+metadata:
+  title: Azure OpenAI Studio for Dialogue Systems
+  description: "In this unit, learners explore the features and capabilities of Azure OpenAI Studio, including model selection, custom training, real-time testing, and multi-channel deployment. They learn how the platform supports scalable, secure development of AI-powered dialogue systems across industries. Keywords: Azure OpenAI Studio, dialogue system development, GPT-4, Codex, real-time AI testing, multi-channel deployment."
+
+  ms.date: 4/28/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+  
+durationInMinutes: 10
+content: |
+  [!include[](includes/2-azure-openai-studio-for-dialogue-systems.md)]

learn-pr/philanthropies/forging-voices-from-data/3-build-and-train-custom-ai-dialogue-models.yml

@@ -0,0 +1,15 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.build-and-train-custom-ai-dialogue-models
+title: Build and train custom AI dialogue models
+metadata:
+  title: Build and Train Custom AI Dialogue Models
+  description: "In this unit, learners walk through the process of building and training AI-powered dialogue models using Azure OpenAI Studio. Topics include setting up an environment, creating deployments, fine-tuning models with custom data, personalizing AI behavior, and iterative testing. Keywords: Azure OpenAI Studio, dialogue model creation, AI fine-tuning, chatbot development, prompt engineering, custom AI systems."
+
+  ms.date: 4/28/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+  
+durationInMinutes: 10
+content: |
+  [!include[](includes/3-build-and-train-custom-ai-dialogue-models.md)]

learn-pr/philanthropies/forging-voices-from-data/4-build-dialogue-systems-that-feel-human.yml

@@ -0,0 +1,15 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.build-dialogue-systems-that-feel-human
+title: Build dialogue systems that feel human
+metadata:
+  title: Build Dialogue Systems That Feel Human
+  description: "In this unit, learners explore advanced techniques for making dialogue systems more human-centered and engaging. Topics include maintaining conversational context, personalizing interactions, incorporating emotional nuance, and using feedback loops to refine user experience. Keywords: AI dialogue systems, personalization, conversational AI, emotional intelligence, context retention, Azure OpenAI Studio."
+
+  ms.date: 4/28/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+  
+durationInMinutes: 10
+content: |
+  [!include[](includes/4-build-dialogue-systems-that-feel-human.md)]

learn-pr/philanthropies/forging-voices-from-data/5-exercise.yml

@@ -0,0 +1,15 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.exercise
+title: Train and customize a dialogue model in Azure OpenAI Studio
+metadata:
+  title: Exercise - Train and Customize a Dialogue Model in Azure OpenAI Studio
+  description: "In this unit, learners practice enhancing a chatbot by uploading a custom dataset, configure system prompts for tone, and simulate conversational memory using Azure OpenAI Studio. They test and refine the model's responses to create a more domain-specific and engaging AI assistant. Keywords: Azure OpenAI Studio, chatbot customization, prompt engineering, fine-tuning AI models, dialogue memory simulation."
+
+  ms.date: 4/28/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+  
+durationInMinutes: 20
+content: |
+  [!include[](includes/5-exercise.md)]

learn-pr/philanthropies/forging-voices-from-data/6-knowledge-check.yml

@@ -0,0 +1,78 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.knowledge-check
+
+title: Knowledge check
+metadata:
+  title: Knowledge Check
+  description: "Knowledge check for the module."
+  ms.date: 3/20/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+
+
+durationInMinutes: 10
+
+
+content: |
+quiz:
+    questions:
+    - content: "What is Azure OpenAI Studio primarily used for?"
+      choices:
+      - content: "To build, customize, and deploy AI-powered language models and dialogue systems"
+        isCorrect: true
+        explanation: "Azure OpenAI Studio provides tools for designing, testing, and refining AI-driven applications using models like GPT."
+      - content: "To design 3D characters for video games"
+        isCorrect: false
+        explanation: "Azure OpenAI Studio is focused on text-based AI, not 3D asset creation."
+      - content: "To manage web hosting and domain registration"
+        isCorrect: false
+        explanation: "Web hosting is unrelated to Azure OpenAI Studio's dialogue system capabilities."
+
+    - content: "What is one key benefit of using Azure for deploying AI-powered chatbots?"
+      choices:
+      - content: "It allows scalable deployments and seamless integration with other Microsoft services"
+        isCorrect: true
+        explanation: "Azure's infrastructure supports enterprise-scale deployments and offers integration with services like Power Automate and Cognitive Services."
+      - content: "It automatically animates characters during dialogue"
+        isCorrect: false
+        explanation: "Azure OpenAI Studio focuses on language models, not animation or motion graphics."
+      - content: "It writes legal contracts from PDF templates"
+        isCorrect: false
+        explanation: "While AI can generate text, legal contract automation is not the platform's core use case."
+
+    - content: "How can users tailor a dialogue model to a specific domain?"
+      choices:
+      - content: "By uploading custom training data such as FAQs, chat logs, or policy documents"
+        isCorrect: true
+        explanation: "Custom data helps fine-tune models so responses are aligned with specific topics, tone, or terminology."
+      - content: "By assigning the chatbot a random name and using default prompts"
+        isCorrect: false
+        explanation: "Personalization and performance come from tailored prompts and training—not randomization."
+      - content: "By selecting a visual theme in the Azure dashboard"
+        isCorrect: false
+        explanation: "Azure OpenAI Studio does not use visual themes—it focuses on natural language input and output."
+
+    - content: "Why is it important to customize system prompts when building a chatbot?"
+      choices:
+      - content: "To shape the assistant's tone, behavior, and response style"
+        isCorrect: true
+        explanation: "System prompts provide instruction to guide the model's personality and communication style."
+      - content: "To allow the model to design its own data sets"
+        isCorrect: false
+        explanation: "The model doesn't create training data on its own—it learns from what you provide."
+      - content: "To automatically translate dialogue into different languages"
+        isCorrect: false
+        explanation: "Prompt customization controls behavior—not translation, which would require additional services."
+
+    - content: "What does real-time testing in Azure OpenAI Studio allow you to do?"
+      choices:
+      - content: "Evaluate and refine chatbot responses before deployment"
+        isCorrect: true
+        explanation: "Real-time testing lets developers simulate interactions, identify gaps, and improve performance."
+      - content: "Create animated facial expressions for each message"
+        isCorrect: false
+        explanation: "Facial animation is not part of Azure OpenAI Studio's functionality."
+      - content: "Publish chatbots directly to the Apple App Store"
+        isCorrect: false
+        explanation: "Azure enables backend deployment—you'll need other platforms for app distribution."

learn-pr/philanthropies/forging-voices-from-data/7-summary.yml

@@ -0,0 +1,15 @@
+### YamlMime:ModuleUnit
+uid: learn.philanthropies.forging-voices-from-data.summary
+title: Summary
+metadata:
+  title: Summary
+  description: "In this unit, learners review the key skills developed for building, training, and refining AI-powered dialogue systems using Azure OpenAI Studio. They summarize the creation process from account setup to prompt customization, and explore techniques for making conversations more human, coherent, and user-centric. Keywords: Azure OpenAI Studio, dialogue system development, AI chatbot fine-tuning, conversational AI, prompt engineering, personalized dialogue systems."
+
+  ms.date: 4/28/2025
+  author: kprks
+  ms.author: kbarreto
+  ms.topic: unit
+  
+durationInMinutes: 5
+content: |
+  [!include[](includes/7-summary.md)]

learn-pr/philanthropies/forging-voices-from-data/includes/1-introduction.md

@@ -0,0 +1,33 @@
+In this session, learners will explore the full lifecycle of designing, customizing, and deploying AI-powered dialogue systems using **Azure OpenAI Studio**. Building on their foundational understanding of conversational AI, participants will learn how to select models, train them with custom data, and refine their outputs through real-time testing.  
+
+The session covers not only the technical steps involved in setting up a dialogue system but also the creative techniques that bring conversations to life—like personalization, tone control, and contextual memory. By the end, learners will have hands-on knowledge of how to create intelligent, human-like assistants that can scale across digital platforms.
+
+## Scenario
+
+Imagine you’ve been asked to develop a virtual assistant for a university’s advising department. The assistant needs to help students find answers about academic policies, course schedules, and campus resources.
+
+Using **Azure OpenAI Studio**, you:
+- Set up a custom instance and select a language model suited to the assistant’s scope.
+- Upload FAQs and policy documents, training the model to reflect the university’s tone and terminology.
+- Customize system prompts to match the university’s professional yet welcoming voice.
+- Monitor interactions as students use the assistant, refining its responses based on feedback to improve clarity and engagement.
+
+With each iteration, you improve the assistant’s ability to serve students effectively.
+
+## What will we be doing?
+
+We’ll walk through the end-to-end process of designing and training a dialogue model in **Azure OpenAI Studio**:
+
+- **Explore**: What is Azure OpenAI Studio, and how can it be used to create powerful dialogue systems?
+- **Select**: How do you choose between GPT-3, GPT-4, and Codex for your use case?
+- **Train**: How can you fine-tune a model using custom datasets and feedback loops?
+- **Customize**: What techniques can make dialogue feel more personal, human, and engaging?
+- **Deploy**: How do you test, evaluate, and improve your assistant across real-world scenarios?
+
+
+## What is the main goal?
+
+By the end of this session, learners will be able to:
+- Build and train custom dialogue systems using **Azure OpenAI Studio**
+- Personalize responses and maintain context across conversations
+- Continuously refine and improve their systems for real-world deployment

learn-pr/philanthropies/forging-voices-from-data/includes/2-azure-openai-studio-for-dialogue-systems.md

@@ -0,0 +1,55 @@
+Azure OpenAI Studio is a cloud-based platform that lets you build, customize, and deploy AI-powered dialogue systems using cutting-edge language models. With its user-friendly tools and connection to Microsoft Azure, it makes creating intelligent, scalable applications easier for both technical and non-technical teams.
+
+## What is Azure OpenAI Studio?
+
+Azure OpenAI Studio gives you an interactive space to work with models like GPT-3, GPT-4, and Codex. You can quickly design, test, and launch AI-powered applications that understand and generate human-like language.
+
+**Key features:**
+- **Prompt playground** — Test prompts live and see how the model responds.
+- **Custom deployments** — Create and manage model endpoints for real apps.
+- **Secure infrastructure** — Built on Azure, offering top-level security and compliance.
+
+**Why use Azure?**
+- **Scalable** — Easily handle thousands of users at once.
+- **Integrated** — Connect smoothly with other Azure services (like Logic Apps or Power Automate).
+- **Real-time testing** — Check and improve model behavior right inside the Studio.
+
+Azure OpenAI Studio helps you move smoothly from idea to production, all in one place.
+
+## What can you do with Azure OpenAI Studio?
+
+Azure OpenAI Studio offers powerful tools to tailor models to your project and audience.
+
+**Model options:**
+- **GPT-3 / GPT-4** — Great for natural conversations, chatbots, and content generation.
+- **Codex** — Perfect for code generation or developer-focused tools.
+
+**Custom training:**
+- Upload your own datasets (like FAQs or customer chats).
+- Fine-tune the model for better, domain-specific results.
+- Use prompt engineering to adjust tone, style, or behavior.
+
+**Easy integrations:**
+- **Azure Cognitive Services** — Add speech-to-text, sentiment analysis, or translation.
+- **Power Automate** — Build workflows triggered by chat inputs.
+- **Logic Apps** — Connect to databases, CRMs, or APIs without heavy coding.
+
+**Deploy anywhere:**
+- Launch chatbots on websites, mobile apps, Microsoft Teams, or social media.
+- Build once, deploy across all platforms for a consistent experience.
+
+## Real-time testing and feedback
+
+Azure OpenAI Studio makes it simple to test your system as you build.
+
+**How to Test:**
+- Use the chat playground to simulate real conversations.
+- Try different inputs and tweak prompts to improve responses.
+- Check tone, clarity, and accuracy on the fly.
+
+**Why It Matters:**
+- Quickly spot gaps or confusing outputs.
+- Refine the dialogue flow without rebuilding or redeploying.
+- Create better user experiences through continuous improvement.
+
+With its real-time tools, Azure OpenAI Studio supports a fast, flexible development loop — helping you craft smarter, more human-like dialogue systems.

learn-pr/philanthropies/forging-voices-from-data/includes/3-build-and-train-custom-ai-dialogue-models.md

@@ -0,0 +1,81 @@
+Creating a custom dialogue system is about more than just picking a model — it’s about careful setup, thoughtful training, and ongoing improvement. In this lesson, you walk through **five essential steps** for building and fine-tuning a functional AI-powered assistant using Azure OpenAI Studio. Each step builds technical skill while helping you shape AI behavior for your unique use case.
+
+## Step 1: Set up your Azure OpenAI Studio account
+
+Before you build, you need the right environment.
+
+1. **Create an Azure account**  
+  Visit [azure.microsoft.com](https://azure.microsoft.com) to sign up or log in. An active subscription is required for Azure OpenAI resources.
+
+2. **Access Azure OpenAI Studio**  
+  Go to [Azure OpenAI Studio](https://oai.azure.com/) and confirm that OpenAI access has been approved for your account or organization.
+
+3. **Set up an OpenAI resource**  
+  In the Azure Portal, create a new OpenAI resource. Choose a region (like East US or West Europe) that aligns with your compliance needs.
+
+**Why it matters:** This setup ensures you’re working in a secure, scalable environment that can handle large language models.
+
+
+## Step 2: Create a new dialogue model
+
+With your setup complete, it’s time to build.
+
+1. **Explore the Azure OpenAI Studio dashboard**  
+  This dashboard gives you access to models, playgrounds, deployments, and prompt configurations.
+
+2. **Start a new project**  
+  Create a new deployment and select the right model (for example, GPT-3 for general tasks or GPT-4 for more advanced dialogue).
+
+3. **Define your model’s purpose**  
+  Clarify the assistant’s role:  
+  Will it help users schedule meetings? Answer policy questions? Act as a digital concierge?
+
+**Why it matters:** A clear scope helps you design focused data inputs and prompt strategies.
+
+## Step 3: Train the model with custom data
+
+Training on your own data fine-tunes the assistant to your needs.
+
+1. **Upload relevant data**  
+  Provide chat logs, FAQs, product details, or service transcripts — real-world examples help the model learn.
+
+2. **Fine-tune with example pairs**  
+  Use prompt engineering or fine-tuning to reinforce the patterns you want. Provide both user input and ideal output examples.
+
+3. **Apply supervised and active learning**  
+  Label outputs, collect feedback, and retrain as needed to improve accuracy over time.
+
+**Why it matters:** This step ensures your assistant isn’t just smart — it’s smart for your audience.
+
+## Step 4: Customize responses and behavior
+
+Shape your assistant’s tone, style, and personality.
+
+1. **Set a system prompt**  
+  Provide instructions that guide all responses, like:  
+  *"You are a friendly assistant who helps answer questions about university admissions."*
+
+2. **Manage context and memory**  
+  Configure how the model handles follow-up questions. Use prompt chaining or summarized context to simulate conversation memory.
+
+3. **Add dynamic variables**  
+  Use placeholders like `{user_name}` or `{appointment_time}` to personalize replies.
+
+**Why it matters:** Customization makes your assistant not just functional, but engaging and on-brand.
+
+## Step 5: Test, evaluate, and iterate
+
+Great dialogue systems are built through continuous refinement.
+
+1. **Run simulated scenarios**  
+  Use the Azure chat playground to test real-world user cases. Identify strong points and gaps.
+
+2. **Collect and analyze feedback**  
+  Monitor interactions to spot confusion or drop-offs. Gather direct feedback when possible.
+
+3. **Refine and redeploy regularly**  
+  Update prompts, retrain on new data, and redeploy improvements to keep your system sharp.
+
+**Why it matters:** Iteration is the key to long-term performance and reliability.
+
+By following these five steps, you move from **concept to custom-built AI assistant** — shaping not just what your system can do, but how well it connects with the people who use it.