AI-powered products started popping up everywhere in 2023. Figma bought Diagram, Framer announced a new AI feature, Canva revealed the new visual worksuite -- changing the way product teams operates.
In 2024, AI is the norm in many product - as well as in designer's workflows. Some feel threatened, others see the opportunity to replace tasks (not jobs!) and improve their creative abilities.
In this article, we will talk about:
- Why is AI key for designers to understand
- How to design with AI: Improve your workflow
- How to design for AI: Learn how to build your own AI product
- Learning in public is your new portfolio
Why is AI key for designers to understand
According to this article, Jakob Nielsen announces AI as the first new UI paradigm in 60 years.
He argues that AI is introducing the third user-interface paradigm in computing history, shifting to a new interaction mechanism where users tell the computer what they want, not how to do it.
While we used to tell the computer what to do, we now tell it which output we want. The machine now runs itself.
This change has an impact on user's expectations, behaviors (how does this work, it's too opaque) but also concerns (Is AI right and objective? how biased is it?) and privacy (who and how is my data used?).
So, if you're looking to become an AI designer, you can have immense leverage by:
- understanding this new user expectation
- new mental models required when designing AI products (intent vs command)
- and fundamentally different user interactions (chats, efficiency)
Given the paradigm shift we experience, you have the opportunity to design with AI to improve your own product design workflow. You can go further and learn the foundational skills to design for AI and build AI applications.
How to design with AI: Improve your workflow
Will AI replace your job?
First, let's adress this fear: will AI replace your job?
According to this article in HBR, AI won't replace entire jobs, but tasks. If most of it is based on repetitive tasks -- then it can be replaced. If part of it is based on critical judgment, reasoning and even empathy -- it can be augmented with AI.
And this is an opportunity for you. If you spend most of your time transcribing research interview notes, yes this can be replaced. However, if you automate this and focus on leading high quality user interviews, then AI will help you focus much more on that part.
AI in itself won't replace your job, but a person using AI definitely will. So, ask yourself the same questions in your product design role?
- First, break down your job into specific tasks . For example, UX researchers perform a lot of research questions, user research interviews, note taking and calls transcriptions, and insight analysis.
- Second, determine whether the task involves the intensive use of language (natural, computational, or mathematical) for completion.
- Third, assess how knowledge is used to perform a particular task. Are its problems ambiguous?
- Once a job has been broken down into its constituent tasks, you can then analyze how generative AI might affect each of those tasks. As a rule of thumb, tasks that entail recurring processes are candidates for full automation with generative AI, while tasks that require creative reasoning, collaboration, and judgment are candidates for augmentation with AI.
- Then, identify which existing tools can help automate or augment some of these tasks you mentionned.
Specific example for UX Researchers
Maureen Herben, a Senior Product Designer, explores how to leverage AI in your UX design process. UX Researchers can use AI to research questions, your planning of research or even double checking research.
- You can use Otter AI to easily record and transcribe user research interviews, summarize and highlight insights.
- You can use Chat GPT to get general inputs for ideation, by getting insights on general design trends (as a starting point). Then get better at asking the right questions.
- You can use MidJourney for vizualisation, but also to get better at generating prompts.
AI Tools examples for each step of your product design workflow
Here is an overview of how AI impacts core product design tasks and specific AI tools that can enhance each aspect of your workflow.
Core Tasks | AI Impact | Key AI Tools |
---|---|---|
User Research | - Analyzing large datasets for insights - Generating user personas - Automating sentiment analysis |
- IBM Watson Analytics - Coco (Figma plugin) - Hotjar AI |
Ideation and Conceptualization | - Generating diverse design concepts - Creating AI-powered mood boards - Data-driven feature prioritization |
- Midjourney - DALL-E - Adobe Firefly |
Wireframing and Prototyping | - Generating wireframes from sketches - Creating interactive prototypes - Suggesting layout improvements |
- Uizard - Sketch2Code - Figma Auto Layout |
Visual Design | - Automated color palette suggestions - AI-powered image editing - Intelligent font pairing |
- Adobe Sensei - Remove.bg - Khroma |
User Testing | - Analyzing user testing sessions - Simulating user interactions - Providing predictive analytics |
- Attention Insight - PlaybookUX - UserTesting AI |
Collaboration | - Automating design handoff - Real-time design suggestions - Translating design specs for developers |
- Zeplin's Project Assistant - InVision Studio - Abstract |
Design System Management | - Updating design components - Identifying inconsistencies - Suggesting system optimizations |
- Figma Variants - Sketch Libraries - Adobe XD Components |
Improve your current design workflow in your job
If you're currently a full-time designer, your best bet is to try and improve your workflow and showcase how impactful you are for your company. It can increase your productivity, your mental health (less stress), and your impact on the business.
A great example is the website you're on right now (Woody) that I manually aggregated data for. In the last few months, I completely automated repetitive tasks using AI, as well as improving my workflow (yes, I believe AI can do a better and quicker job than me on repetitive tasks). For example:
- All job descriptions are cleaned-up using a specific prompt
- Company description are auto-generated using a specific prompt
It seems like nothing, but I saved up to 3 hours per day. Now, I can focus on tasks that are more creative, and business-oriented.
If you think like an entrepreneur, you will want to automate, streamline, reduce all parts of the business that is repetitive. If you have this mindset, you will constantly remove blockers for companies you work with and prove that you're invaluable.
How to design for AI: Learn how to build your own AI product
Don't Take Courses, Build your Own AI product
If you're curious about AI, and actually learn the product and technical challenges it entails - it has never been cheaper and "easier" to build your own thing. This is a great way to learn new skills, showcase them effectively (and attract hiring managers) and join an AI-first company.
Even if you don't build everything, understanding how building and training your own model, prompt engineering, and how it can improve effciency in business can make a big difference for you into landing AI Designers jobs.
If you are building a small AI tool to learn basics, I recommend you start solving a problem that you have as a product designer. You can follow exactly the same steps to identify how specific tasks of your job can be replaced by AI, and build an AI solution for it.
Start with the tedious work that junior product designer have to do. Remember when we chatted about tasks vs jobs. List out all the tasks thats are repetitive and require an intensive use of language. For example: taking notes of user interviews, designing low-fi prototypes, writing summaries of your design decisions for your colleagues and boss.
Try and solve the problem with NO AI at all. This forces you to break the problem down into lots of discrete pieces that you can write normal, traditional code for and see how far you can solve this.
Don't use general LLMs, but instead design and train your own model. It sounds hard, but according to builder io, training your own AI model is a lot easier than you probably think with only basic development skills or no-code using Pecan.
Clean up your data using Google's Vertex and test your AI
Build up your web applications using code or tools like Bubble
Test it out for yourself!
Don't Take Courses, Build your Own AI product (no-code friendly)
Bonus for product designers who don't dabble with code, give the no-code x AI a try. This is also a great way to grasp technical concepts and increase your changes to join AI teams. Here is a video you might learn from:
Learning in public is your new portfolio
To leverage your learning and land a job for AI designers, I recommend that you start documenting and sharing your learnings in public. It is a unique way to connect with people who share similar interests.
For example, you can:
- Share on Linkedin how you much time you saved doing research using tools like Otter.ai
- If you're more into videos, showcase how you created your first AI app using no-code tools and what you learnt from this
- If you have a blog in your portfolio (and like writing more), describe how you started using predictive analytics to test new designs based on historical user data using DataRobot.
And you won't have to "create your portfolio", but simply share your learnings. The new way of hiring is based on sharing real-life projects.
Hire AI Designers
Post you job on Woody to hire AI designers.