Well-Architected AI workload assessment

The Azure Well-Architected AI workload assessment is a review tool that you can use to self-assess the readiness of your AI workload in production. Running an AI workload on Azure can be a complex process. The assessment is organized so that you can methodically check your workload's alignment with the best practices of the Azure Well-Architected Framework pillars.

For best results, the team that completes the assessment should be well versed in the architecture of your workload. That team should also have a strong understanding of cloud principles and patterns. These roles include but aren't limited to cloud architects, operators, and DevOps engineers.

The assessment is a set of questions that are based on the AI workload design areas. It's a method for checking the foundational design choices of your workload's architecture and your end-to-end operational approach.

These questions are designed to help you benchmark your workload's maturity and alignment with the recommended approach for operating an AI workload. The outcome of the assessment is a set of technical recommendations and documentation that provides guidance about how to implement a highly reliable solution on Azure.

AI assessment tool

Next step

See the following reference architecture, which describes design choices for a production-ready implementation:

• Baseline OpenAI end-to-end chat reference architecture