Generative Design: AI as a Creative Partner in Prototyping
Dr Asma Qureshi
•
September 24, 2025
.png)
•
In recent years, one of the most profound shifts in my workflow has come from the rise of generative design.
What once took hours of careful work in Figma or Sketch, building wireframes, exploring layouts, or mocking up interfaces, can now be jump-started with AI. By simply describing or sketching an idea, I can generate instant design variations that spark creativity and accelerate the process. Instead of starting from a blank canvas, I begin with a range of possibilities, which frees me to focus more on refining, testing, and bringing the most promising ideas to life.
Let me share a personal anecdote: during a consulting project, I sketched a rough mobile app interface on paper, consisting of nothing more than some boxes and scribbled text, to show a basic layout. I then fed a photo of this sketch into Uizard, an AI-powered design tool. In less than a minute, Uizard transformed my hand-drawn outline into a digital prototype with styled components.
It wasn’t production-ready art by any means, but it was a surprisingly coherent UI draft, complete with proper alignment and placeholder content. I sat back and thought, “This felt like science fiction.” What used to take an entire afternoon with a design team happened in moments. Tools like Uizard leverage generative AI to transform sketches or text prompts into wireframes and high-fidelity mockups, significantly reducing the time it takes to go from concept to a testable design.
Even the mainstream tools I use every day are incorporating AI-assistive features. Figma, for example, has introduced AI features that act like an intelligent design assistant. I can ask Figma's AI to generate multiple variations of a button or suggest alternatives for an icon, and it will produce options almost instantly. They say a picture is worth a thousand words; in this case, an AI-generated picture can be worth a thousand ideas.
However, in adopting generative design, I’ve also recognised its limits. AI is great at generating average solutions very quickly. Still, it doesn’t inherently understand the subtleties of my product’s brand, the emotional tone we need to strike, or the nuanced needs of my users. For instance, an AI might churn out ten layout variations for a dashboard. Still, only a human designer (in collaboration with the team) can judge which one truly aligns with the product vision or is the most accessible and culturally appropriate.
I often treat AI-generated designs as starting points for my own work. They give me something concrete to critique and build upon. In fact, a common practice on my team is to take an AI-generated wireframe and then spend our time in design reviews focusing on refining details, text hierarchy, spacing, colour choices, and so on, rather than starting from a blank canvas. This accelerates our process while keeping us in control of the creative decisions.
In short, AI has become my creative partner, not a replacement for creativity. It's there to beat the blank page syndrome and handle the grunt work of producing iterations. I still recall a recent project where we needed several form screen variations to conduct an A/B test. Instead of manually designing each one, I described the form's purpose to an AI tool and received several reasonable layouts.
We quickly identified one that seemed most promising, and I then polished it by applying our design system and tweaking the copy. The result was something that met our standards, delivered in a fraction of the time it would have taken otherwise. That experience reinforced for me that generative design isn't about letting go of the wheel; it’s about driving faster with a turbo boost, all while keeping my hands ready to steer.
Askable is the modern research platform for velocity-obsessed teams.
Let's chatWith all these advancements, I constantly remind myself (and my team) that just because we can do something with AI doesn’t always mean we should. The integration of AI into UX entails significant ethical responsibilities. In my role, I've made it a priority to address issues of bias, privacy, transparency, and human oversight at every step.
AI systems learn from data, and if that data isn't carefully vetted, the AI's outputs can reflect and even amplify societal biases. This is a genuine concern in UX, because a biased design or recommendation can alienate entire groups of users. I recall an incident (fortunately caught in testing) where an AI-driven feature started suggesting different credit card offers to users based on inferred demographics, effectively reinforcing stereotypes in who was offered what. It was a stark lesson. Now I emphasise feeding our AI tools with diverse, representative data and continuously monitoring their outputs for fairness.
If we ask an AI to generate user personas or content suggestions, we double-check that it’s not skewing to just one archetype. Inclusivity isn’t something the AI will naturally handle; it’s something we have to ensure through careful oversight and dataset curation. In practice, this may involve augmenting training data with inputs from underrepresented user groups or implementing rules within the algorithm to mitigate bias. It’s an ongoing effort, but it’s absolutely vital for ethical UX design.
.png)
Many AI-driven UX improvements rely on user data, including clicks, scrolls, purchases, and even location or sensor data. That data is precious, and how we handle it can make or break user trust. I take privacy very seriously. When we use AI analytics tools, we ensure that user data is anonymised and only collect what is genuinely necessary. With regulations like GDPR (in the EU) and CCPA (in California), there are legal requirements to meet, but to me, it's about more than compliance; it's about respect.
I also advocate for transparency with users: if an AI feature uses personal data to personalise the experience, we consider telling the user why we're asking for specific information and how it benefits them. For instance, a mobile app might say, "Enable location to get restaurant recommendations in your area." Clarity like that goes a long way. And of course, data security is non-negotiable when AI crunches user data. We ensure it’s protected and stored securely, because any breach could severely erode the trust we’ve worked hard to build.
Get a sneak peak into the product, and everything Askable can do for you.
Contact salesOn that note, one thing I often emphasise when speaking to peers or mentees is that no matter how advanced AI becomes, there are uniquely human qualities it cannot replicate. AI doesn’t feel joy or frustration; it can't truly empathise with a user's context or imagine what it's like to be in someone's shoes. Those empathetic insights often guide my most crucial design decisions, especially in crafting tone, emotional resonance, or delight in an experience. Likewise, human creativity can leap beyond existing patterns in a way AI (which learns from the past) simply can’t.
The creative leap that led to, say, the first iPhone interface was not something AI could have predicted by analysing past phones. It took human vision. Remembering this keeps me and my team humble about the role of AI. As a colleague of mine wisely put it, AI is a brilliant co-pilot, but it still needs a pilot. I couldn't agree more. In my work, critical thinking and ethical judgment remain fundamental human duties, and AI's contributions, however impressive, are designed to amplify our human abilities, not replace them.
Standing at the intersection of design and technology, I’m both excited and sobered by what’s on the horizon. The near future of UX design will likely be defined by even more seamless integration of AI, and it’s my job to stay ahead of the curve while keeping my perspective grounded.
One trend already emerging is real-time adaptation. We touched on personalisation, but I foresee AI driving interfaces that adapt not just to a user's past behaviour, but to what's happening in the very moment. Think of a software platform that can sense when a user hesitates or appears confused, perhaps through cursor movements or pause times, and then instantly adjusts by offering a helpful tip or simplifying the options on screen. This kind of context-aware, real-time personalisation is becoming feasible with advanced machine learning models and faster computing.
Design That Feels Like a Conversation
In fact, we're already seeing early versions of this. Some modern UX tools can analyse user sentiment and engagement on the fly, giving designers instant feedback and allowing us to fine-tune experiences as people interact. It's like having a live conversation with your user base, where the design can respond in the moment. Done well, this could make interfaces feel more alive and accommodating. Done poorly, of course, it could also overwhelm or annoy users, so we'll have to iterate carefully and always ask whether a real-time change truly benefits the user.
Another exciting frontier is the fusion of AI with augmented reality (AR) and virtual reality (VR). As AR/VR technologies become more mainstream (with devices like the latest AR glasses and VR headsets), AI will be crucial in making those experiences rich and responsive. I've been following how AI improves aspects such as 3D rendering and environmental interactions. For example, AI-driven enhancements can make virtual environments and characters more realistic and intelligent, reacting to user actions in believable ways. This has enormous implications for training simulations, gaming, and even virtual collaboration tools.
Additionally, AI will further streamline our workflow and automate design processes. We've already discussed generative design tools; looking ahead, I expect the entire UX lifecycle to be augmented. Routine tasks, such as quality assurance testing, can be automated primarily by AI scanning our designs for issues (some of this is already being implemented; for instance, AI can check colour contrast against accessibility standards).
User research ops could also be automated end-to-end: imagine an AI scheduling test sessions with representative users, conducting the sessions via an AI persona, and then summarising findings in a report, all while you sleep. While that scenario might be a few years away, we're headed in that direction. In development, there are AI systems that can translate visual designs into code. It's plausible that soon designers and developers will work even closer, with AI acting as the translator to ensure the implemented product matches the design perfectly.
Sources:
Looking Ahead
In conclusion, I remain optimistic and pragmatic about the role of AI in UX. Every day, I'm amazed by what technology can do, analysing in seconds what used to take weeks, generating ideas I wouldn't have thought of, and personalising experiences in wonderfully human-like ways. It's transforming our field, mainly for the better, by augmenting our capabilities as researchers and designers. But I also remind myself that our role as human-centred designers is more important than ever. AI lacks empathy, intuition, and the capacity to bear responsibility if something goes wrong. Those are our domains.
The magic happens when we blend the two: using AI's power and efficiency with our human insight and care. In my experience, when a design team strikes the right balance, the result is the best of both worlds – user experiences that are deeply personal, delightfully efficient, and ethically sound. As we move forward, my advice to fellow design professionals and executives is to embrace AI's potential with enthusiasm but do so with eyes open and hands firmly on the wheel. AI will undoubtedly shape the future of UX, but it's up to us to shape the future of AI in a way that truly serves people. And that, I believe, is a fascinating design challenge.
.png)
.png)