The past couple weeks I've been resharpening my UI design skills to design the next version of Semanttic (something to share soon!). It's also given me a chance to explore first-hand how to utilize AI tools in the typical day-to-day UI production work. Generative AI is pretty good at ideation, but this is a tiny part of what we do in UI work. So where else can AI be helpful?
The following are things I've been using for. I've been using Perplexity a lot more recently as I find it just gives me higher-quality answers, but your mileage may vary. Whether you're using Perplexity, ChatGPT, or something else, I think the following can work for you so give it a try and let me know
The mainstream generative AI tools are all pretty decent at quick feedback. They can help you identify areas for improvement, suggest design principles to follow, and offer insights on how to enhance user experience. It doesn't replace having a trusted colleague provide feedback, but I've found it help see the work in a different way and it's very quick. It will also include usability and interaction design best practices, which is helpful in anchoring your thought process during this time.
I was a big fan of Kuler back in the day, which was absorbed into Adobe. Today, Adobe has some of the best tools for browsing or creating color palettes. This hunting process can take awhile though and where I've been able to use AI is to describe the feel I'm looking for, or a certain mood, and have it make suggestions. I like that it tries to provide some context for why it's suggesting what it has, as well as the hex values that I can plug in right way.
Typography, like color, is another area where I can end up burning a lot of time. My background is Interaction Design and I'm not the strongest graphic designer. My process has typically been to try lots of different things and mostly second-guess myself a lot until something eventually clicks; or, I get inspired by a different product and try and apply the same type they use. Perplexity was pretty good at making some recommendations and evaluating the choices I had made. It's important to consider AI as a quick and easy source of feedback, not the ultimate source of truth. You still need to have conviction to make the final decision, but I found I was able to have something close to a design conversation at the exact moment I needed it.
I used this a lot. I stopped using production design tools in my day-to-day work over 10 years ago, when more of my time was spent on the people and business parts of my job. While the principles of great design are universal and unchanging (or very slow to change and evolve), the tools change much faster. I know my around Figma but I'm not an advanced user, and Perplexity was brilliant at helping me level up much faster than I would have other wise. For example, I couldn't figure out how to fill a shape with a pattern. It's something I know how to do in Illustrator but could not figure out how to do in Figma. Perplexity gave me the perfect answer. There have been maybe a dozen or so examples like this during this redesign work I'm doing.
I've been thinking about how designers will change how they work because of Generative AI, when so little of what we do is pure ideation. What we need are tools that reduce the friction in day-to-day production work, or eliminate it all together.
Some questions that I've been thinking about:
- How much of this needs to be built into one super tool vs a handful of specialty tools?
- What are effective ways of providing the right context for LLMs?
- Where do we integrate chat into a design-specific tool, like Semanttic, without it becoming too much cognitive overhead for people?
I encourage all designers to start experimenting with simple, practical ways of utilising AI in their daily work. I think it can be a fun way to change up your day and keep you creatively fresh, and you may find in a very short amount of time that you're able to do things in minutes that take you hours or days today.