How incorporating AI-based tools into your design process makes you a more efficient designer
How to use AI-based tools effectively in your design process
Interviewing design candidates showed me the subtle effects and advantages Artificial Intelligence (AI) has had on the design field.
Yes, massive failures are happening with AI right now, and yes, there is controversy about companies faking AI. But don't let all the controversial buzz surrounding AI distract you from one fundamental truth: Designers who use AI will be more efficient than those who don't.
I realized this when a design candidate provided a comprehensive topic overview for a whiteboard exercise they knew nothing about 72 hours ago.
Here are some ways I've learned to use AI-based tools throughout my design process and how you can learn to use them efficiently as well. But before that, I need to emphasize my view on AI.
The fundamental issue for designers with AI-based tools is trust
AI will not replace designers simply because the fundamental issue with it is trust.
If you rely solely on AI to do your design work, you are being dumb. Not only might this convince your organization to lay you off (since you're showing them anyone can use AI to do design), but this action leads to a loss of trust with designers.
If you mistakenly design a product no one wants, based on faulty AI-based user research, you'll either be fired, or your creative freedom (i.e., design process) will be restricted. So, the best way to think about AI-based tools is like a clumsy personal assistant: it can help you but needs guidance.
While generative AI like Midjourney can be interesting, I use AI-based tools to automate my design process's tedious (and less critical) parts. Doing so allows me to focus on the parts of design that require deep thinking to solve complex problems.
Learning to delegate these tasks, along with several name-brand AI tools, allows people to not only trust my process more but also allows me to be much more efficient.
Here are the AI tools I use the most.
Summarize meeting minutes for alignment with Microsoft Copilot
I used to capture meeting minutes, and it sucked.
Capturing meeting minutes is outside the job description for Design roles, but when you're a one-person UX team or working with people who change their minds all the time, tracking the big decisions made and the next steps is critical.
So, I used to assemble e-mail summaries and meeting minutes to turn verbal approval into written approval. Assembling these used to take forever, but it's become much easier with Microsoft Copilot.
Microsoft Teams previously had the functionality to record and transcribe meetings, and they've used AI to summarize what happened quickly and the significant points raised.
Of course, this is not a signal to stop taking notes. One of the rough patches I've often encountered is that Copilot skips over certain points that you might find important. For example, Copilot might skip a quick conversation between your Product Manager and you about a UX-related question if most of the discussion is with the whole team about something else.
However, it's still a way to quickly summarize a particular meeting, especially if you want to provide that to someone who couldn't attend. Doing so helps everyone be on the same page and communicates your design intent and previous decisions around that step.
These sorts of summaries are often one of the strongest use cases for AI, which leads to a more controversial topic: summarizing user research.
Getting another research/design perspective with ChatGPT
Do not let ChatGPT do user research for you. Doing so is one of the surest ways to get fired, as ChatGPT is not a trustworthy source for research and design recommendations.
However, ChatGPT can provide interrater reliability, a fancy academic term for comparing notes with others.
When you're part of a large design team and multiple UX professionals participate in user research sessions, it's easy to debrief and see what people think are the most important ideas, their main takeaways, and more.
But bouncing ideas off people is hard for a UX team of one. Whether discussing user research findings or design solutions, having someone weigh in and provide feedback can help immensely. This is where ChatGPT can help.
For example, once you analyze your user research, you can have ChatGPT summarize the results and compare them to see what you have in common and what seems to be a major point.
However, you must do the work first. Whether it's putting together the user research or creating design solutions, you will run into far fewer headaches if you do the work first and then check it against AI.
There are two reasons for this. The first is that ChatGPT excels at example-based thinking: if you can provide an example with your prompt, your results will always be better. However, the second is more important: you know why you made certain design decisions or think certain findings are important.
Lastly, one of the things I never really see mentioned but is incredibly useful is the ability to generate templates.
Figjam and the ability to generate design artifact templates
One of the most common AI-based tools I've used is FigJam's Generative Template tool.
Typing in "I want you to generate X" can quickly generate a template that meets your specifications. This is where you might say, "How does that save any time compared to using a template?" The answer, as it turns out, is in personalization.
Imagine you wanted to create a design persona with very specific categories and looks. Until now, you would search for it on Google/Figma Community, which may yield many user-generated results that all look slightly different.
As a result, you’d have to adjust some fields here and there or move some things around. Each persona template has a few things you might want to change, and selecting which persona out of a list of hundreds also takes time.
This time spent adds up to more than you might think. But what if you typed exactly what you wanted into FigJam instead? For example, "I want a design persona with (A, B, C, D, and E) fields, a round circle for an image, a title and subtitle, and an extensive background section?"
While it’s a different level of fidelity, FigJam can generate these templates with this filled-in data (and these sections) to be personalized to your needs. Rather than searching for a "close enough" template and adjusting it to your needs, this ties into one of AI's greatest strengths: personalization.
One of the great things AI can offer is the ability to create a template solely for your project, your needs, mine, and no one else's.
AI will not replace designers, but Designers who know AI will
Jakob Nielsen provided that quote, which I've seen reflected when hiring a new designer.
I've seen design candidates who provided comprehensive responses to complex questions in a really short timeframe that would have required one of two things in the past:
Pulling several all-nighters or working a long time on this project
Getting help from several other designers to help put everything together
However, the advent of AI has provided many tools to make us more efficient. Does that mean it's perfect or that we should entrust the future of Design to AI? Of course not.
However, it's indisputable that AI allows designers to work more efficiently. More importantly, it allows us to be efficient with things that we probably find boring or struggle with, giving us more time for deep design thinking and iteration.
This is especially useful if you're a UX team of one (or a Senior Designer). When you often struggle to find time to do deep design work with many meetings, the last thing you need to eat up that time is tedious work.
So, even as the hype around AI fails, you should realize that it's a tool to help designers work more efficiently around tasks they might otherwise find boring. Understanding that mindset can help you effectively use the Power of AI.
I’ve revamped my Maven course to teach Data Informed Design. If you want to learn this valuable skill, consider joining the waitlist.
Kai Wong is a Senior Product Designer and creator of the Data and Design newsletter. His book, Data-Informed UX Design, provides 21 small changes you can make to your design process to leverage the power of data and design.