AWS Well-Architected Assistant

• 

  AWS Well-Architected AI Assistant

Inspiration

Designing AWS architectures and operating workloads according to design principles and best practices is not an easy task. AWS Well-architected Framework provides great support, but also requires a lot of knowledge. The AWS Well-Architected Tool implements a structured approach to evaluate system designs, but also requires an experienced architect to get the most out of it and to avoid stepping in the process of just check boxes.

It would be great if people can describe there most important system requirements and the current design in natural language and let an AI assistant do the review.

What it does

The app generates a review for each of the six AWS Well-Architected Framework pillars based on the user's input (system requirements and design, described in natural language). The pillar reviews include "Positives", "Potential issues" and "Suggestions". In addition, an overall improvement plan is provided as a summary. Users can update their input at any time to refine the review iteratively. Users can also chat with the assistant about the review, e.g. asking questions about risks or how to star t with improvements.

How I built it

partyrock.aws generative app builder, leveraging text and chat widgets and some prompt engineering.

Challenges I ran into

Generating an review for one pillar at a time and not include aspects of other pillars, especially for Sustainability.

Accomplishments that I am proud of

A useful tool that generates insights and makes a well-architected review more efficient.

What I learned

Generative AI / Large-Language Models can be used for analytical tasks.

What's next for AWS Well-Architected Assistant

Supporting AWS Well-Architected Lenses like Serverless, SaaS and IoT.

Built With

• partyrock.aws