Brief Table of Contents (Not Yet Final)
Preface
Introduction
1. Introduction to Generative AI and the Azure OpenAI Service
What’s Artificial Intelligence
Current Level of AI Adoption
The Many Technologies of AI
Typical AI Use Cases
Types of AI Learning Approaches
About Generative AI
Main Capabilities
Relevant Industry Actors
The Key Role of Foundation Models
Road to Artificial General Intelligence (?)
Microsoft, OpenAI, and the Azure OpenAI Service
The Rise of the AI Copilots
Azure OpenAI Capabilities and Use Cases
LLM Tokens as the New Unit of Measure
Conclusion
2. Designing Cloud-Native Architectures for Generative AI
Modernizing Applications to Make Them Generative AI-Ready
Cloud Native Development
Microservice-based Apps and Containers
Serverless Workflows
Azure-based Web Development and CI/CD
Understanding the Azure Portal
General Azure OpenAI Considerations
Available Azure OpenAI Models
Architectural Elements for Generative AI Systems
Conclusion
3. Implementing Cloud-Native Generative AI with Azure OpenAI
Defining the Knowledge Scope of Azure OpenAI-enabled Apps
Generative AI Modelling with Azure OpenAI
Azure OpenAI Service Building Blocks
Visual Interfaces: Azure OpenAI Studio and Playground
Playground #1
Playground #2
Playground #3
Playground #4
Deployment Interfaces: WebApps and Power Virtual Agents
Development Interfaces: APIs and SDKs
Interoperability Features: Function Calling and “JSONization”
Potential Implementation Approaches
Basic Azure ChatGPT instance
Minimal customization with one or few-shot learning
Fine-tuned GPT models
Embedding-based grounding
Document indexation / retrieval-based grounding
Other grounding techniques
Approach Comparison and Final Recommendation
AI Performance Evaluation Methods
Conclusion
4. Additional Cloud and AI capabilities
Plugins
LLM Development, Orchestration, and Integration
LangChain
Semantic Kernel
Bot Framework
Power Platform, Power Virtual Agents, and AI Builder
Databases / Vector Stores
Vector Search from Azure Cognitive Search
Vector Search from CosmosDB
Redis Databases on Azure
Other Relevant Databases (Including Open Source)
Others Microsoft Building Blocks for Generative AI
Azure AI Document Intelligence (Formerly Azure Form Recognizer) for OCR
Ongoing Microsoft Open Source and Research Projects
Conclusion
5. Operationalizing Generative AI Implementations
The Art of Prompt Engineering
Generative AI and LLMOps
Prompt Flow and Azure ML
Securing LLMs
Managing Privacy and Compliance
Responsible AI and new regulations
Relevant Regulatory Context for Generative AI Systems
Company-level AI Governance Resources
Technical-level Responsible AI tools
Conclusion
6. Elaborating Generative AI Business Cases
Pre-Mortem or What to Consider Before Implementing a Generative AI Project
Defining Implementation Approach, Resources, and Project Roadmap
Defining Project Workstreams
Identifying Required Resources
Estimating Duration and Effort
Creating a “Living” Roadmap
Creating Usage Scenarios
Calculating Cost and Potential ROI
Conclusion