Designing UX using AI

I enjoy building quality user interfaces which are usable and attractive, and it’s great having AI tools to assist with this. My main approach for designing UX is to use the integrated AI agent within my IDE to generate lots of UX options which can be displayed within the app using dummy data.

My prompt is something like this:

I want to build a contact page to display the user’s address book. Brainstorm several UX options and implement them all, using dummy data to create a realistic design. Add a temporary dropdown to allow me to select between these designs.

The AI will then build a bunch of options within your app which you can evaluate. Bear in mind that some may be quite poor! However, some may have elements which appeal to you. Give the AI agent feedback by telling it which elements of which designs you liked, and ask it to make another design combining these elements.

Often I might go through this brainstorm & feedback loop several times as I get towards a solution I am happy with.

There are various online AI tools which can produce entire UX designs from a well written prompt. These are great for building out a concept, but a key advantage of designing within your application is that you can test out the design as if you were a user, and see it interacting with the rest of your app (or at least with your dummy data). Seeing the design of a UI component in context helps you decide if it’s worth keeping.

As AI can produce amazing results, why not just let AI make the design without human input? While you can do this, you can often end up with a subpar design. On the other hand, a good product engineer using AI can produce high quality and useful designs relatively quickly. The key skills of the product engineer here are to write specific prompts, have an eye for design, and give good feedback to the AI agent.

TLDR: You can build amazing user interfaces quickly using AI - as long as you have a good eye for design and give accurate feedback.