Free AI Design Toolkit by UX Woman

Get access to our free AI design toolkit to learn the principles of AI product design. We include case studies from Netflix, Spotify, Google, Amazon and more within interactive learning experiences, videos, written case studies, and FigJam templates.

0-1 AI Product Design Innovation

“AI” is officially the buzzword I got tired of hearing about in 2023 and 2024. But all that hype got to me and I became an avid ChatGPT user last year.I fell in love. Collaborating with ChatGPT was like chatting with a friendly co-worker throughout the workday. ChatGPT was great at making suggestions and helping me with research. 


But it also had its limits, and I realized I was not going to be replaced anytime soon as a creator, researcher, and designer! I started seeing ads for AI apps pop up all over my Instagram feed, sometimes making ridiculous claims, like “Get AI to run your business” or “Get a personal AI CFO”. It got me thinking…

  • 👉 How can we tell the difference between BS AI apps vs. the good ones? AI isn’t going anywhere. In fact, the AI market will be worth about $244 billion by the end of 2025 and is projected to grow to $827 billion by 2030. 

    👉 The field of AI product design is already booming and will continue to boom. I was chatting with tech friends who currently work at Amazon and Square and we were discussing how AI tools are becoming more prevalent at work, while designing AI products is now becoming the norm. 

    👉 But here’s the secret: you can start designing AI products right now by leveraging large language models (LLMs) like ChatGPT and Gemini. Now and in the future, ChatGPT is going to mimic Apple’s app store, with thousands of apps leveraging the LLM capabilities provided. 


    But that doesn’t necessarily mean that every single AI product is going to be great. In fact, we are already seeing a lot of sh**ty AI products and AI-generated content spamming our social media and news feeds.

Delusional by Design Challenge: Redesign Popular AI Features from Spotify & Google Maps

Consider AI design systems, heuristics, and atomic design in these 15-minute ideation challenges! Let’s get started.


  • What AI features are incorporated into Google Maps? More than you might think! Google is working on improving a whole portfolio of AI features when it comes to their Google Maps product, one of the most sophisticated mapping systems in the world, with vast amounts of data mapping literally every inch (or almost every inch) of this world. 

    Google Maps has invested heavily into AI summaries of a variety of locations, from coffee shops to restaurants to national parks, museums, hiking trails, and scenic spots. These summaries are also available in search results, different locations prioritized by AI depending on the search.

    AI is also heavily being used to plan routes and actually “see” places before you get there via 3D immersive views of a variety of locations. But while all this fancy AI tech is great and continues to be splashed across news headlines on a regular basis, is it actually helping?

    Is Google Maps actually a better experience because of AI? Hmm, for us the answer is yes and no. AI features in Google Maps have a lot of potential to make our lives easier, but sometimes they miss the mark. 

    What’s an example of ‘good’ Google Maps AI? We think it’s pretty great that Google Maps can analyze our routes in real-time and make suggestions to switch to a faster route depending on traffic patterns. Google will usually interrupt your drive to ask you if you want to switch to the faster route, giving you a few seconds to make the switch while trying to minimize notifications and interruptions while you drive. 

    What’s an example of ‘bad’ or ‘could be improved’ Google Maps AI? We think that the Google AI search recommendations could be improved. Google can rely on a vast data set of locations all over the world to help us discover new places…however, that recommendation system doesn’t seem to understand what we want as humans.

    This is where user research, UX strategy, and product strategy are crucial so that we can understand user perspectives and provide them with the best types of recommendations. We tend to be in different moods when we travel- sometimes we just want to get to our destination fast. Other times we want to explore local restaurants. And other times we want to sightsee as we stroll about aimlessly within a city. 

    When I was in Rome, I used Google Maps to go from monument to monument, sightseeing by choosing my own way and wandering about the city for hours. I got annoyed whenever Google recommended that I stop at a commercial store in between sights. I really was not in the mood to shop, and the algorithm should have understood this via my searches.

    In this way, Google Maps was trying to prioritize consumerism, spending, and advertising prime mainstream store locations rather than just letting me enjoy local sights and local mom-and-pop shops. I ended up switching off the maps for a bit, just to get lost in the city for a while. 

    So how might we leverage Google Maps’ vast repository of locations to suit user needs when route planning? Watch our 15-min Redesign Google Maps AI challenge! 

    BTW, if you want to dive deeper into AI design heuristics while coming up with your own ideas for redesigning Spotify’s AI, we include our AI design heuristics FigJam toolkit to help you design AI features and products. Get access by subscribing to the UX Woman Substack ($6.67/month). Once you subscribe, you will gain access to all templates, FigJam boards, our VIP UX job board and case studies that we only include for VIP subscribers at the bottom of each of these emails.

    Our AI Design heuristics FigJam template includes:

Redesign Netflix AI in 15 Mins- A FigJam Session

  • The very nature of AI is unpredictable, to the point where even AI’s engineers, creators, and designers cannot fully explain how complex AI outputs are created. 

    We have literally created something whose capabilities are often beyond our control, and it is human nature to fear uncertainty and that lack of control, especially when designing AI features and products.

    So, the best brands design for uncertainty. They create flexible and breathable designs that ensure there’s room for whatever outputs the AI generates. 

    You can watch our Redesign Netflix’s AI Features in 15 minutes video here for a full breakdown in FigJam ideating on wireframes and new ideas. 

    What do I mean by flexible and breathable designs? I’m talking about creating gallery templates where multiple AI outputs can be featured and prioritized in a variety of manners so that users can make the right choices based on their needs.

    In this case, gallery designs should not overwhelm users, but should provide them with just enough context they need to take action and make a decision. As designers, we might want to consider reducing the number of choices, even if that means increasing AI error/uncertainty rates…see more in the videos above. It’s always about striking a delicate balance.

    If you want to dive into AI Product Design, consider taking our 0-1 AI Product Design Innovation sprint course. We cover:

    👉 What is AI and how it works. We need to have a basic understanding of AI so that we can design great AI products.

    👉 How AI and people interact. How might we design for handoffs between users and AI, and vice versa? What are the inputs and outputs?

    👉 AI product design heuristics. What are best practices to consider when designing AI product interfaces? 

    👉 AI prototyping. How might we scope high-impact, low-effort features? How might we make our product more sophisticated over time as we scale it?

    👉AI control vs. user choice. How might we strike a balance between AI automation and user control so that outcomes are beneficial for users?

    👉Designing for trust. How might we foster user trust in our AI systems, without leading to over-reliance?

    Watch our free AI Product Design course preview here to get an understanding of why you should be thinking about AI product design today. This is a beginner friendly UX course. 

    If you would like a deeper dive into designing AI products and features for uncertainty, get the full case study here. In our case study we cover:

    👉 Designing for AI uncertainty and error. How does AI work and what causes uncertainty and error in AI outputs? What are ways to ethically design AI features and products to consider error rates, while still giving users choices and agency over their tasks at hand? 

    👉 2 ways AI can be deceptive. What are some ways that AI lies to us? And why do designers sometimes have to design with deception? What’s the difference between good deception and bad deception?

Redesign Google Search AI in 15 Mins- A FigJam Session

  • Just wanted to remind you that UX design is supposed to be fun and playful! It can be draining to look at tech news, apply to UX jobs, and stress out about career paths.

    Relax. Take a break! Come redesign Google Search with us for fun for 15 mins. What’s this challenge about? 

    Not taking ourselves too seriously and embracing a playful, entrepreneurial mindset when redesigning Google Search’s AI features, while also considering their business strategy, monetization, and an ethical user experience.

    Google has been prioritizing its AI Overview feature a LOT recently, sometimes even prioritizing it over sponsored content, which is really interesting to see.

    It’s also interesting to see a variety of different search “templates” or search layouts depending on what is searched. In the examples above, different layouts are formatted depending on the search type: entry level ux jobs, ux design, and ux courses. 

    Which brings us to AI design. AI outputs can be unpredictable, which is why as AI designers, we have to learn to design for uncertainty and provide a variety of options that AI can use to prioritize different types of information.

    For example, depending on the search a Google layout could either prioritize sponsored ad content, an AI overview, the most popular search results from websites, social media community posts from Reddit or Quora, or Tiktok & YouTube videos. 

    Essentially, the AI algorithm decides what design element gets prioritized depending on the search and their understanding of the user. From a UX design systems perspective, this gives us an opportunity to create a variety of “atoms” and “molecules” that the algorithm can take and re-prioritize depending on the search needs, in order to create whole Google search templates. Watch the full video to get an understanding of this introduction to design systems! 

    Also, what might the AI Google Search of the future look like, 10-20 years from today? Certainly it would not be the same.

    In our VIP subscriber version of this email ($6.67/month), we dive into the future of Google Search, 10-20 years from now and discuss the long-term consequences of the AI features we design today, including previews from our AI Product Innovation course. Subscribe to gain access to:

    If you need a 1:1 consultation on how to design an AI-based product or feature, consider taking our online AI Product Innovation course, which includes a mentorship consultation with Samaya.