In earlier days, UI/UX Designers has to translate every single detail given by stakeholders and project owners. We have to give our interpretation of the end product as soon as possible. We tend to do our research, look at references on the internet and improvise it to our design. The process can take a long time since the more complex the app, the longer time it takes to create a wireframe. UI/UX Designers had to carefully design features, managing it's placement, making sure it aligns with stakeholder and user's needs, etc.

Now with Google Stitch (previously Galileo AI), everyone can generate the general framework of the design to make it easier for user to translate the needs or general image of the app. Google Stitch provides us with quick wireframe that we can use as a reference for our app. It indeed needs a specific prompt because shorter prompt will make 

However, things aren't like how we thought it be. With ChatGPT, whatever prompt we type in, most of the time they knew what our intentions and gave us the information we needed. Remember the AI video of Will Smith eating spaghetti in 2022? I believe Google Stitch are currently at that phase of development. To prove my opinion, I decided to create a simple wireframe of a weather tracking app. Keep in mind that I'm using the standard mode of Stitch.

My design here included the weather (described with logo), temperature, location, prediction over few upcoming hours, and some metrics like humidity, air quality, UV Index, and Average Temperature. Here's my prompt for weather tracking app.

While not perfect, I expect my prompt to be similar, and here's the result I got from my first prompt

As we can see, Stitch generated my design in their own style. The hourly forecast is quite informative, the icons were clear enough although the icon at 1 PM is quite unique, but it's taking a lot of space horizontally. There's also no additional metrics like temperature tracker, prediction for upcoming hours, function to change location etc.

On my attempt to generate newer design, I then try to add some corrections to some feature by asking Stitch to change the weather prediction into a graph model. I did this to make small adjustments and prevent from making significant changes.

This time, the design had some improvement. It shows in 12 hour there will be 2° increase of temperature. However, weather prediction turns into a graph but without any metric (what the scale goes up or goes down means ). Moreover, we didn't know what 21° means (is it in celsius, fahrenheit, etc). To fix this, I decided for my next prompt would first to correct the scale on the graph.

I decided to test for the last time with this prompt

As we can see, there's not much change on the interface. The information however change from temperature of 21° into 70° and it still indicates sunny, which is we still didn't know what the temperature metrics are.

I do think it's gonna take a while before we can "humanize" this AI. But like that Will Smith video, in only 3 years, we can develop Google Veo 3, a video generating tool powered by AI, and the result has been on the internet for a while. AI for UI UX design are progressing massively and it's only a matter of time before we can see competitors on the market. I do believe that AI is a tool that can help to create better product rather than replacing the workers. For that, I am optimistic enough to see what Google Stitch can develop in the future.