Will AI fundamentally change product design? When my third-grader created a sophisticated thank you card in minutes using Canva’s AI tool, it was clear that AI is accelerating the democratization of design. My third-grader didn’t need to study typography, have an expensive software license, or watch any tutorials on YouTube. All she needed was a clear idea of what she wanted and a place to enter some text.
AI is changing the work we do as designers and how we work. This is the first post in a three part series exploring the potential transformative impact of artificial intelligence on product design.
Trends and AI impact on design – Exploring AI-driven trends and their impact on design
The evolution of design practice – What is this means for design and designers
Future-proofing designers – The essential skills and mindset designers need in the age of AI
Transforming Tedious Tasks
I started out in design doing lots of repetitive work like updating text, swapping out images, and changing fonts. Many designers are frustrated that they spend most of their day doing repetitive work and ensuring their engineering team has all of the flows they need and wish they had more time for strategic and exploratory work. AI tools can do the work they don’t want too but it may mean a company needs fewer designers:
Adobe Sensei GenAI can generate images, use text effects, and generate lots of options for a creative director’s review.
Nutella used an algorithm to generate seven million unique jars and they sold out in a single month.
AirBnb created a prototype that matches a sketch to the design system and generates code.
ChatGPT can save time by creating product images and text so you don’t have to write it yourself or wait for marketing
Many companies have invested in design systems to ensure consistency with reusable components, guidelines, and documentation. The system houses the decisions of design experts that are available to everyone. A well-defined design system and good AI tools can speed up the creation of mockups substantially.
Unfortunately, research teams have been hit hard by the layoffs in the last year and AI makes it possible for a small team to do more. A survey from User Interviews shows that AI tools are popular for automation even though there are concerns that AI will introduce bias or create inaccurate analysis.
About a third of researchers are using AI for document signature collection (34.5%), scheduling (32.5%), and screening applicants (31.6%)
47.8% use AI for transcription.
45.5% use AI to help with writing reports
Democratizing design by simplifying the tools
The worst insult a designer used to get years ago was “I could have made that myself.” Now you really can make just about anything a designer can make with an AI enabled tool. These tools are unlikely to create something great and unique but you can make something acceptable.
Even the tools of the professional designer are more accessible than ever. When I started, there were a lot of buttons you could press and little documentation on how they worked. For many years, designers have been really fortunate to have loads of great and free resources to help from knowledge bases to YouTube tutorials.
The tools have gotten a lot more friendly too. When we switched to Figma about two years ago, we taught product managers and marketers how to modify designs and update copy in less than an hour. Don’t worry, Figma has great version control, so we can fix things when they get broken.
Instead of just knowing the tools, I also need to know how best to give directions and feedback in a text box.
When you take on your first design leadership role the hardest part is learning how to guide someone instead of doing the work yourself and now we all need to be able to do it.
Getting prototypes to customers faster than ever
Our goal is to test and learn as early and often as we can. AI-powered prototyping tools like Uizard can help designers create interactive prototypes in hours (not weeks), allowing for faster validation of product concepts and iterations. Galileo AI takes natural language prompts and generates UI designs for a specific product. This allows us to experiment and to really invest in the design once we know what it needs to do.
Empowering teams with easily accessible data and insights
The CEO of Pendo Todd Olson shared on The Product Experience Podcast that they get about 10,000 product requests a month. That is a great job for AI. That data can point you in the right direction for interviews, and then you can go back to those customers to test the new solution.
A hot topic in research is whether you should use AI tools for research. I can see the value in using something like SyntheticUsers to do quick tests and get a quick perspective before doing your testing with customers. I see it as a tool in the box and not a replacement.
Many organizations are investing in analytics tools to better understand user behavior.
Often, getting to the data is challenging and requires a lot of knowledge of the tool and how the product was tagged. Busy designers and PMs don’t have time for this, being able to ask a well-formed question and get data aggregated from different tools using AI would dramatically increase the use of data in product and design decisions.
Creating personal and accessible experiences
Netflix can’t only generate images; they use an algorithm to pick the artwork that best highlights the film for you. We can offer customers many paths and options based on what we know about you and others like you.
Oftentimes, product teams focus on creating features that will serve big groups of customers, and if you aren’t in one of those groups, the features you need won’t make it in. However, an AI integration can unlock data, content, or functions that would never make sense for a product team to build.
Designers often want to improve accessibility but struggle to get the necessary investments and time to make improvements. New tools are offering designers many ways to serve the diverse needs of customers.
Google Vision API uses neural networks to help visually impaired users identify images and assess their safety.
CAPTCHA is annoying for sighted people but really difficult for someone who is visually impaired; facial recognition would be much easier.
Google is working on an AI to convert lip movements into text in real-time to help users with accurate text summaries.
AI-powered navigation systems suggest accessible routes and provide information about nearby accessible facilities, improving mobility.
Innovating with information
With AI, businesses can generate forecasts based on vast amounts of data, recognizing patterns and making predictions, which can be essential for optimizing prices, increasing profit, and making informed decisions. We can extract this data and present the right data at the right time in the way clients need it.
One interesting thing we’re seeing now is that AI is not a great tool for predictions for a professional audience. Professionals often want to know what the methodology was and what peers were chosen and why. This may be overcome eventually as tools get better at explainability and professionals get more comfortable with substituting their judgment for the AI’s.
Partnering with the expert you need in real-time
Designers are using AI tools to act as their strategic partner by:
Getting feedback on a presentation deck
Running competitive analysis on products or companies
Applying strategic frameworks like the sample attribute map from Anne Morris and Frances Frei or the Kano model
Imagining what an expert on strategy like Michael Porter would say about the topic
AI tools are useful in generating and pressure-testing ideas. “Since the AI doesn’t have an ego, you can use LLMs and the visual AI models to produce ideas your team never would have. These ideas can be easily discarded, but can spark other ideas that have real potential. In the grind of efficiency and factory-style UX, this is an efficient way to step out of that and play with new ideas,” says product and technology expert Scott Varho.
You can’t always get on a call with a colleague to get input on your work. AI can review a document or a design and offer feedback that can help a designer nudge their work a little further and catch mistakes.
I do worry that creativity may take a hit in a AI driven world, I find a lot of what comes out of these models are flat and generic.
Navigating the ethical landscape
We’ve already seen AI supercharging fraud, generating misinformation and disinformation, and producing biased results. Designers are beginning to think about what tools, practices, and governance are needed to prevent these challenges. Some organizations are doing important work in this area:
So what does this mean for designers? In Part 2, I’ll explore how this changes the role of designers, the principles that guide us, and the approaches we use. In Part 3, I’ll explore how this changes design education and training.
I’m learning a lot of this as I go and I’m thankful for all the people helping me along. Please share your thoughts and questions so we can learn and explore together.
Image: I usually make all the images on my site because it’s fun for me but for an AI series I had to go with DALL-E