Share via

Solving Artificial Intelligence problems

  • Article

The Artificial Intelligence (AI) Playbook provides enterprise software engineers with solutions, capabilities, and code developed to solve real-world AI problems. Everything in the playbook is validated.

This playbook provides AI solutions focused in the following areas:

  • Pre-trained ML models, such as Large Language Models (LLM), and Azure AI Services (at times referred to as Model as a Service or MaaS).
  • Custom Machine Learning Models - that are developed with Azure tools such as Azure Machine Learning (AML), Databricks, Synapse and Fabric.
  • Software as a Service (SaaS) solutions, such as Microsoft's PowerVirtualAgents.

Some good places to start:

Tech Topic Getting Started Solution
Generative AI Working with LLMs
Azure ML Getting Started with Azure ML Automated and Monitored Training Pipelines

About the AI Playbook

The AI Playbook aims to accelerate real-world application development by following good engineering practices, such as:

  • Improving application design and developer productivity by sharing code and knowledge developed by experts for Microsoft customers.
  • Using automation to make repetitive tasks faster, more reliable, and auditable
  • Making application deployments and operations secure and observable.
  • Securing applications by following security best practices at all stages of the engineering lifecycle.
  • Writing portable solutions that run in multiple locations, including edge devices, on-premises data centers, the Microsoft cloud, and competitor's clouds.

Integrated solutions

Playbook solutions span multiple Microsoft products and services and focus on creating integrated end-to-end solutions often using a range of open-source software libraries.

Proven with real customers

The guidance, documentation, and code in this playbook are based on real-world engineering projects Microsoft has completed with customers. While the documentation and code has been generalized to remove confidential information, it is real code that has been deployed by Microsoft customers. The playbook shares real-world approaches and learnings that have worked for real projects.