Skip to content

Commit 903eb62

Browse files
authored
Merge pull request #51048 from MicrosoftDocs/NEW-architecture-of-thought
New architecture of thought -> main -- ASAP
2 parents de558e5 + ed57668 commit 903eb62

16 files changed

+509
-0
lines changed
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.introduction
3+
title: Introduction
4+
metadata:
5+
title: Introduction
6+
description: "In this unit, learners explore the foundational concepts of AI-powered dialogue systems. They learn how Azure OpenAI Studio supports the creation of chatbots and virtual assistants, focus on dialogue architecture, conversational flows, and real-world applications. Keywords: Azure OpenAI Studio, dialogue systems, conversational AI, chatbot development, AI-powered assistants."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 5
14+
content: |
15+
[!include[](includes/1-introduction.md)]
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.ai-powered-dialogue-systems
3+
title: AI-powered dialogue systems
4+
metadata:
5+
title: AI-Powered Dialogue Systems
6+
description: "In this unit, learners explore what AI-powered dialogue systems are, how they work, and why they are critical to modern digital experiences. They study real-world examples across industries and understand how conversational AI improves user engagement, task automation, and service availability. Keywords: AI dialogue systems, conversational AI, natural language processing, customer support chatbots, virtual assistants."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 10
14+
content: |
15+
[!include[](includes/2-ai-powered-dialogue-systems.md)]
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.components-and-architecture-of-dialogue-systems
3+
title: Components and architecture of dialogue systems
4+
metadata:
5+
title: Components and Architecture of Dialogue Systems
6+
description: "In this unit, learners explore the core components that make up an AI-powered dialogue system, including user interfaces, natural language processing engines, dialogue management, knowledge bases, and response generation. They learn how each layer works together to create natural, intelligent conversations. Keywords: dialogue system architecture, AI conversation design, NLP, dialogue management, response generation."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 10
14+
content: |
15+
[!include[](includes/3-components-and-architecture-of-dialogue-systems.md)]
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.overview-of-ai-models-used-in-dialogue-systems
3+
title: Overview of AI models used in dialogue systems
4+
metadata:
5+
title: Overview of AI Models Used in Dialogue Systems
6+
description: "In this unit, learners explore the AI models that power dialogue systems, including natural language processing models like GPT-3 and GPT-4. They examine how conversational AI frameworks create human-like interactions through language understanding, response generation, and reinforcement learning. Keywords: conversational AI, GPT models, natural language processing, AI dialogue systems, reinforcement learning."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 10
14+
content: |
15+
[!include[](includes/4-overview-of-ai-models-used-in-dialogue-systems.md)]
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.azure-openai-studio-overview
3+
title: Azure OpenAI Studio overview
4+
metadata:
5+
title: Azure OpenAI Studio Overview
6+
description: "In this unit, learners explore Azure OpenAI Studio, a cloud-based platform for building, testing, and deploying AI-powered applications using models like GPT-3 and GPT-4. They learn how the platform supports prompt engineering, secure deployments, and scalable integration into real-world applications. Keywords: Azure OpenAI Studio, GPT-3, GPT-4, AI application development, AI prompt engineering."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 10
14+
content: |
15+
[!include[](includes/5-azure-openai-studio-overview.md)]
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.exercise
3+
title: Build a basic AI-powered chatbot with Azure OpenAI Studio
4+
metadata:
5+
title: Exercise - Build a Basic AI-Powered Chatbot with Azure OpenAI Studio
6+
description: "In this hands-on activity, learners use Azure OpenAI Studio to build a simple AI-powered chatbot capable of handling basic customer support queries. They practice designing system prompts, test conversational flows, and refine chatbot behavior through prompt iteration. Keywords: Azure OpenAI Studio, AI chatbot, prompt engineering, chatbot development, conversational AI."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 15
14+
content: |
15+
[!include[](includes/6-exercise.md)]
Lines changed: 109 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,109 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.knowledge-check
3+
4+
title: Knowledge check
5+
metadata:
6+
title: Knowledge Check
7+
description: "Knowledge check for the module."
8+
ms.date: 3/20/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
###########################################################################
13+
###
14+
### If your content is related to a product or service, apply one value from the either the ms.prod allowlist
15+
### or the ms.service allowlist. You can’t use both ms.prod and ms.service.
16+
###
17+
### If your content isn't about a product or service, you can omit both ms.prod and ms.service.
18+
###
19+
### The list of approved ms.prod values is here: https://review.learn.microsoft.com/help/platform/metadata-taxonomies?branch=main#msprod
20+
### The list of approved ms.service values is here: https://review.learn.microsoft.com/help/platform/metadata-taxonomies?branch=main#msservice
21+
### If you need to request new values, follow the process here: https://review.learn.microsoft.com/en-us/help/platform/metadata-allowlist-requests?branch=main
22+
23+
durationInMinutes: 10
24+
###########################################################################
25+
###
26+
### General guidance (https://review.docs.microsoft.com/learn-docs/docs/id-guidance-knowledge-check)
27+
###  - Questions are complete sentences ending with a question mark
28+
###  - No true/false questions
29+
###  - 3 answers per question
30+
###  - All answers about the same length
31+
###  - Numeric answers listed in sorted order
32+
###  - No "All of the above" and/or "None of the above" as answer choices
33+
###  - No "Not" or "Except" in questions
34+
###  - No second person ("you") in the questions or answers
35+
###  - Provide a meaningful explanation for both correct and incorrect answers
36+
###
37+
### Question content requirements:
38+
### - Write 5 questions
39+
### - Questions 1,2 must test this Learning Objective: "Describe how (attributes) of (product) work to (solve problem)"
40+
### Guidance: These two questions can be short, no need for a long scenario to analyze. Test if they understand how the product works.
41+
### Example: "What differentiates an action from a control action in a Logic App?"
42+
### - Questions 3,4,5 must test this Learning Objective: "Evaluate whether (product) is appropriate to (general product use case)"
43+
### Guidance: Use scenario questions that ask the learner to analyze a situation with the "when to use" criteria presented in the module.
44+
### Example: "A financial company is building a system to let brokers trade financial instruments. The system must monitor market conditions, detect changes, and execute trades. It will need to handle a large volume of transactions quickly. The faster it completes trades, the more of an advantage the company will have over its competitors. Which requirement of this system would be difficult for Azure Logic Apps to satisfy?"
45+
###
46+
###########################################################################
47+
48+
content: |
49+
quiz:
50+
questions:
51+
- content: "What is the primary purpose of using DALL·E in interactive storytelling?"
52+
choices:
53+
- content: "To generate character portraits based on descriptive text prompts"
54+
isCorrect: true
55+
explanation: "DALL·E transforms written character descriptions into visual portraits, helping storytellers visualize and develop their cast."
56+
- content: "To create branching dialogue trees for NPCs"
57+
isCorrect: false
58+
explanation: "DALL·E is used for image generation—dialogue trees are created using language models like GPT."
59+
- content: "To code decision logic in interactive games"
60+
isCorrect: false
61+
explanation: "Coding logic is not part of DALL·E's functionality—it is focused on visual generation."
62+
63+
- content: "How does Azure OpenAI assist with dialogue generation in storytelling?"
64+
choices:
65+
- content: "By using AI models to generate responsive, branching dialogue based on character traits and narrative context"
66+
isCorrect: true
67+
explanation: "Azure OpenAI's language models generate text-based dialogue that can vary based on character personalities and story flow."
68+
- content: "By animating character portraits with real-time voice syncing"
69+
isCorrect: false
70+
explanation: "Azure OpenAI focuses on language, not animation or voice syncing."
71+
- content: "By importing stock dialogue clips from a Microsoft media library"
72+
isCorrect: false
73+
explanation: "Azure OpenAI generates new, custom text—it doesn't use a stock media library."
74+
75+
- content: "What is a key benefit of using AI to generate dialogue trees?"
76+
choices:
77+
- content: "It enables faster creation of varied, character-driven conversations"
78+
isCorrect: true
79+
explanation: "AI speeds up the creation of multi-branch dialogues, while allowing characters to reflect distinct personalities and tones."
80+
- content: "It eliminates the need for story structure or player choice"
81+
isCorrect: false
82+
explanation: "AI supports but does not replace narrative planning or interactivity design."
83+
- content: "It automatically records and voices over all NPC dialogue"
84+
isCorrect: false
85+
explanation: "Voiceover is not handled by Azure OpenAI—it provides text, not audio."
86+
87+
- content: "Why is human oversight important when using AI-generated dialogue?"
88+
choices:
89+
- content: "To ensure tone, style, and narrative consistency align with the writer's vision"
90+
isCorrect: true
91+
explanation: "Human review helps shape AI output to match the world, emotion, and tone of the overall story."
92+
- content: "Because AI tools require manual drawing input for each character"
93+
isCorrect: false
94+
explanation: "AI does not require manual drawing—DALL·E generates images from text."
95+
- content: "To translate dialogue into different programming languages"
96+
isCorrect: false
97+
explanation: "Oversight is for narrative alignment, not code translation."
98+
99+
- content: "How can AI-generated dialogue improve player immersion?"
100+
choices:
101+
- content: "By enabling responsive conversations that reflect player decisions and character traits"
102+
isCorrect: true
103+
explanation: "AI can create dialogue paths that adjust to player behavior, increasing emotional depth and engagement."
104+
- content: "By creating automated background music based on mood"
105+
isCorrect: false
106+
explanation: "AI-generated dialogue is focused on language, not audio scoring."
107+
- content: "By writing entire game scripts with no user input"
108+
isCorrect: false
109+
explanation: "While AI can support scripting, storytelling requires ongoing human input and creativity."
Lines changed: 15 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,15 @@
1+
### YamlMime:ModuleUnit
2+
uid: learn.philanthropies.architecture-of-thought.summary
3+
title: Summary
4+
metadata:
5+
title: Summary
6+
description: "In this unit, learners review the key concepts of AI-powered dialogue systems, including their architecture, the role of models like GPT-3 and GPT-4, and how Azure OpenAI Studio supports prototyping and deployment. They reflect on the essential skills needed to build scalable, natural conversational AI experiences. Keywords: dialogue systems, Azure OpenAI Studio, GPT models, conversational AI, AI chatbot development."
7+
8+
ms.date: 4/28/2025
9+
author: kprks
10+
ms.author: kbarreto
11+
ms.topic: unit
12+
13+
durationInMinutes: 5
14+
content: |
15+
[!include[](includes/8-summary.md)]
Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
In this session, learners explore the foundational concepts and real-world applications of AI-powered dialogue systems. With a focus on Azure OpenAI Studio, participants dive into how intelligent agents like chatbots and virtual assistants are built, customized, and deployed to support natural, effective conversations. The session introduces key components of dialogue architecture and shows how AI models automate support, streamline communication, and provide engaging user experiences across industries. Through guided exploration and collaborative activities, learners begin crafting their own dialogue-driven applications using Microsoft tools, setting the stage for deeper hands-on development in later modules.
2+
3+
## Scenario
4+
5+
Imagine you’re designing a customer support chatbot for a startup that wants to offer 24/7 assistance to users across its website and mobile app. The team wants the bot to handle FAQs, triage requests, and hand off complex issues to live agents—all while sounding friendly and professional. You use Azure OpenAI Studio to design a conversational flow, selecting a language model and customizing prompts that reflect the brand’s voice. By integrating sample intents, you shape a bot that greets users, interprets questions, and generates relevant responses. With each iteration, you refine the flow and test outputs, making sure the system handles varied inputs gracefully. The result is a responsive, scalable assistant that saves the company time, reduces support costs, and improves customer satisfaction.
6+
7+
## What will we be doing?
8+
9+
We explore the role and functionality of AI-powered dialogue systems and apply Microsoft tools to start developing them:
10+
11+
* **Define**: What are dialogue systems and where do they appear?
12+
* **Explore**: What makes up a dialogue system’s architecture and workflow?
13+
* **Model**: What kinds of AI models power these interactions?
14+
* **Evaluate**: How do dialogue systems improve efficiency and user experience?
15+
* **Test**: How can Azure OpenAI Studio be used to start prototyping a chatbot?
16+
* **Apply**: How can learners shape dialogue flows that address real-world challenges?
17+
18+
## What is the main goal?
19+
20+
By the end of this session, learners understand the purpose and structure of AI-powered dialogue systems, explore industry examples, and use Azure OpenAI Studio to create intelligent conversational experiences.
Lines changed: 70 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,70 @@
1+
AI-powered dialogue systems transform the way humans interact with technology. These systems enable natural, human-like conversations between users and machines, offering intuitive access to services, support, and content. This lesson explores what these systems are, where they appear, and why they matter in today's digital landscape.
2+
3+
## What is an AI-powered dialogue system?
4+
5+
An AI-powered dialogue system is a software application designed to simulate conversation with users through text or voice. Unlike simple rule-based bots that follow fixed scripts, these systems use artificial intelligence—particularly machine learning and natural language processing (NLP)—to understand language input, manage dialogue flow, and generate meaningful responses in real time.
6+
7+
### Key characteristics
8+
9+
- **Natural language understanding (NLU)** – The system interprets and processes user input in conversational language.
10+
- **Dynamic response generation** – The system generates responses in context, adapting to the conversation flow.
11+
- **Context awareness** – The system remembers prior interactions within a session to maintain continuity.
12+
- **Multi-turn interaction** – The system supports back-and-forth exchanges instead of one-question-one-answer formats.
13+
14+
These capabilities allow dialogue systems to go beyond automated scripts—they become interactive agents that help users solve problems, find information, or co-create content.
15+
16+
## Real-world examples of AI dialogue systems
17+
18+
AI-powered dialogue systems appear across industries to improve customer experience, reduce operational costs, and deliver round-the-clock service. Here are some examples:
19+
20+
### Customer support chatbots
21+
22+
- Companies like Microsoft use chatbots to handle large volumes of routine inquiries.
23+
- These bots provide quick answers to common questions (e.g., order status, password resets).
24+
- They escalate complex cases to human agents when necessary.
25+
26+
### Virtual assistants
27+
28+
- Devices like Microsoft Cortana use voice-based dialogue systems to interact with users.
29+
- They perform tasks like setting reminders, controlling smart home devices, or playing music based on voice commands.
30+
- They learn from user preferences over time to deliver more personalized responses.
31+
32+
### Healthcare assistants
33+
34+
- AI chatbots help with symptom checking, appointment scheduling, and medication reminders.
35+
- They provide basic triage support before human intervention becomes necessary.
36+
37+
### Education and creative applications
38+
39+
- Language learning apps use dialogue systems to help students practice conversations.
40+
- AI storytelling tools let users co-create characters and narratives by chatting with virtual personas.
41+
42+
These examples show the versatility of dialogue systems in both practical and creative domains.
43+
44+
## Role and importance of dialogue systems
45+
46+
AI-powered dialogue systems are not just a trend—they play a key part in the digital transformation happening across industries. Their value comes from their ability to engage users naturally and deliver support efficiently.
47+
48+
### Enhancing user experience
49+
50+
- They provide a more human, conversational way to interact with technology.
51+
- They reduce friction in digital interactions by letting users speak or type as they normally would.
52+
53+
### Automating tasks
54+
55+
- They free up human workers from repetitive, low-complexity tasks.
56+
- They streamline workflows like data collection, appointment setting, or basic troubleshooting.
57+
58+
### Supporting 24/7 availability
59+
60+
- Unlike human teams, dialogue systems stay available any time of day.
61+
- They ensure users can access help when they need it—especially valuable in global or always-online services.
62+
63+
### Reducing operational costs
64+
65+
- Businesses scale their support systems without hiring large teams.
66+
- They reduce call center volume and let human agents focus on higher-impact work.
67+
68+
AI-powered dialogue systems simulate natural conversations using technologies like NLP, machine learning, and dynamic response generation. These systems are used across industries from customer support to healthcare, and they play an essential role in creating scalable, user-friendly, and cost-effective digital experiences.
69+
70+
By understanding what these systems do, how they work, and why they matter, learners see the potential of conversational AI—not only as a tool for automation, but also as a way to design interactions that feel more human.

0 commit comments

Comments
 (0)