Design Thinking in an AI World

There is a strange thing happening in design right now. The tools are getting better at creating interfaces faster than ever, while the industry is still trying to figure out what that means for the people who design them.

A few years ago, being able to take an idea and turn it into a polished interface required a fairly long process. You needed to understand the problem, work through the user experience, create wireframes, build visual designs, prototype the interactions, hand everything off to development, and then spend time making sure the final product actually matched the intent. Today, AI can help with nearly every part of that process. It can generate layouts, write copy, create components, produce code, suggest interactions, create images, and turn an idea into something functional in a fraction of the time and to be clear, I think that’s a very good thing!

I use AI in my own design and development workflow, and I would be lying if I said it has not changed the way I work. It has made me faster, allowed me to explore more ideas, and made the distance between having an idea and actually building it much smaller. As someone who works across UX/UI design and frontend development, that is incredibly useful.

At the same time, it has made me more aware of something I have always believed about design: knowing how to produce an interface is not the same thing as knowing how to design an experience.

That distinction matters more now than it ever has.

When Everyone Can Make Something That Looks Good

One of the biggest changes AI has brought to the design industry is the accessibility of production. You no longer need years of experience with Figma, Photoshop, HTML, CSS, or JavaScript to create something that looks reasonably polished. Someone can describe a website to an AI model and have a functioning version of it in minutes. They can ask for a dashboard, a SaaS application, a marketing site, or a mobile interface and receive something that, at first glance, looks surprisingly convincing to the untrained eye.

That changes what it means to stand out as a designer.

Visual polish is still important. Typography matters. Layout matters. Color, spacing, motion, accessibility, and visual hierarchy all matter. I am not arguing that those skills are becoming irrelevant. What is changing is how difficult it is to produce something that appears polished.

When the tools become better at producing the surface of a design, the deeper parts of design become more important.

This is where I think having a strong understanding of UX principles gives me an advantage.

I am not just looking at whether something looks good. I am looking at whether the information is organized in a way that makes sense, whether the interaction matches the user’s expectations, whether the interface is creating unnecessary cognitive load, whether the hierarchy reflects what the user actually needs to know, whether the system provides enough feedback, and whether the experience still works when the user does something that was not part of the ideal happy path.

Those are not problems that disappear because an AI model can generate a beautiful interface.

In fact, AI can sometimes make those problems harder to notice because the result looks so convincing. I think that is one of the biggest traps we are going to see over the next few years. We are going to have more and more digital products that look like they were designed well without necessarily being designed well.

There is a difference!

UX Is More Than Knowing the Rules

I have always found UX principles more useful when they are treated as tools for thinking rather than rules to memorize.

A designer can know the definition of cognitive load and still create an interface that overwhelms people. They can know about Hick’s Law and still give users too many choices. They can understand information architecture and still organize a website around the company’s internal structure instead of the way customers actually think about the business.

Knowing the principles is only the beginning. The real value comes from being able to recognize when and why they matter.

That is something I pay attention to constantly when I design.

If I am working on a dashboard, I am thinking about what the user needs to understand immediately and what can wait. If I am designing a form, I am thinking about the information the user has available when they reach each field, how errors are communicated, and whether I am asking for information that is actually necessary.

If I am designing a marketing website, I am thinking about how someone who knows nothing about the company is going to understand what it does, whether the page gives them enough context to make a decision, and where the experience should take them next.

Those decisions are not really about making screens, they are about understanding people. That is the part of UX that becomes particularly valuable when AI enters the workflow.

AI Makes My Process Faster, Not Less Human

I do not use AI because I want to remove myself from the design process. I use it because I want to spend less of my time doing things that machines are now very good at and more of my time making decisions that require context and judgment that bring a big return on investment for the business and the users alike.

I can explore several layout directions without spending an entire afternoon manually creating each one. I can move from a design concept into functional frontend code much earlier. I can experiment with interactions that might have previously been too expensive to prototype. I can use AI to help me investigate an unfamiliar technical problem instead of spending hours searching through documentation before I can even begin testing an idea.

But the more I use these tools, the more important it becomes for me to understand what I am looking at.

AI is very good at giving you something plausible. Plausible is not necessarily good!

It might generate a navigation pattern that looks familiar but does not fit the content. It might create a dashboard with all of the right components while completely misunderstanding which information should have the highest visual priority. It might build a responsive layout that technically adapts to a smaller screen while making the actual interaction frustrating on mobile.

If I do not understand UX, I may not recognize those problems.

If I do understand UX, I can use the AI output as a starting point and then make better decisions about what needs to change.

That is the difference between using AI to design and using AI as part of a design process and knowing that difference can easily make or break a business, especially a business in it’s infancy!

The Cost of Getting UX Wrong

There is another side of this conversation that I think gets overlooked, especially when we talk about startups and small businesses. For a small company, the appeal of AI is obvious, budgets are limited and teams are small. There may not be money to hire a UX designer, a UI designer, a frontend developer, a copywriter, a researcher, and a strategist.

AI appears to offer a way to compress all of those roles into a handful of tools and sometimes it can.

But there is a difference between reducing the cost of production and reducing the cost of a bad decision. I think this is where businesses need to be particularly careful.

It is very easy to look at the current AI landscape and think, “Why would I spend thousands of dollars on design when I can just use AI to build it?”

The problem is that building something is often not the expensive part. Building the wrong thing is.

A startup can save money by skipping UX research. It can use AI to generate its website. It can generate the interface, write the copy, produce the code, and launch the product without ever really stopping to ask whether the experience makes sense.

It might look like an incredibly efficient process from the outside.

Then users struggle to understand the product. Conversion rates are lower than expected. Customers cannot find what they need. People abandon onboarding. The information architecture starts breaking as the company adds features. The marketing website does not communicate enough trust. Accessibility problems appear. Developers have to rebuild components because the original structure was never designed to scale.

Now the company is paying someone to fix the things it thought it had saved money on. That is why I think the real value of UX is often misunderstood.

Good UX is not simply an expense associated with making something prettier or easier to use. It is a way of reducing risk.

It helps a business identify problems before they become expensive problems. It helps determine what actually needs to be built. It helps prioritize what matters to the user and what can wait. It creates structure before development begins. It makes decisions intentional instead of reactive.

For a large company, getting a design decision wrong can be expensive. For a startup with ten employees and a limited runway, it can be existential. That is why I would actually argue that small businesses and startups may have even more to gain from strong UX thinking than large organizations do. They simply cannot afford to spend six months building something users do not understand.

The cheapest design decision is rarely the one with the lowest price tag. It is the decision that prevents an expensive mistake later.

The Designer’s Role Is Changing

I think there has been a lot of anxiety around whether AI is going to replace designers and I understand where that anxiety comes from, especially when you see how quickly these tools are improving. But I think the more interesting question is what parts of a designer’s job are actually changing.

Some forms of production are absolutely becoming faster. There is no reason to pretend otherwise. Creating variations, generating initial concepts, writing basic code, producing assets, documenting components, and translating ideas into working prototypes can all happen much faster with AI assistance.

But design has never just been production. There is a significant difference between creating a screen and deciding what the screen needs to accomplish. There is a difference between generating a component and understanding where that component belongs in an experience. There is a difference between producing a working website and understanding whether the website is actually helping someone accomplish what they came there to do.

As AI takes on more of the production work, I think designers will increasingly have to operate at that higher level. We will need to become better at defining problems, understanding users, establishing constraints, evaluating solutions, testing assumptions, and knowing when something that looks impressive is actually making the experience worse.

That is why I think designers should be actively deepening their understanding of UX right now. Not because we need to go backward and become more theoretical. The opposite.

We need to understand the fundamentals so well that they become instinctive. Learn cognitive load, learn information architecture, learn accessibility, learn mental models. Learn affordances and feedback, learn progressive disclosure, learn error prevention, learn visual hierarchy, learn responsive design as an experience problem rather than simply a collection of breakpoints and then bring AI into that knowledge.

Let it challenge your ideas. Let it generate alternatives. Let it help you prototype. Let it write the first version of the code. Let it investigate technical problems.

Let it make you faster, but do not outsource your ability to determine whether the answer is actually good.

Don’t Compete With AI at What It Does Best

I do not think designers should compete with AI at the things AI is becoming exceptionally good at. If AI can generate ten interface concepts in the time it used to take me to create one, I do not want to compete with it by trying to become even faster at producing interface concepts.

I want to become better at deciding which concept deserves to exist.

I want to become better at understanding why something works.

I want to become better at recognizing when a familiar pattern is appropriate and when it is simply familiar.

I want to become better at understanding users, business constraints, accessibility, content, technology, psychology, and the messy reality that exists outside of the ideal user journey. That is where I think designers can create real separation.

The designers who understand UX deeply and know how to use AI as a force multiplier are going to be able to do things that were previously difficult to do within normal timelines and budgets. They can explore more possibilities without getting stuck in production. They can move between design and development without treating them as completely separate disciplines. They can prototype earlier, test ideas sooner, investigate technical constraints earlier, and make more informed decisions throughout the process.

That is increasingly how I think about my own work.

My value is not that I can sit in Figma and manually produce every screen faster than everyone else.

My value is that I can understand the problem, determine what the experience needs to accomplish, design the system around those requirements, use AI to accelerate the parts that make sense to accelerate, and then recognize when the result needs to be changed.

That requires UX knowledge, visual design, technical understanding, business awareness, and judgment. The tools are changing quickly, but those underlying skills are still useful. In fact, I think they become more valuable.

When Production Becomes Cheap, Judgment Becomes Valuable

AI is going to make production cheaper. It is going to make experimentation cheaper. It is going to make prototyping cheaper. It is going to make development more accessible. And when something becomes cheaper to produce, more of it gets produced.

We are going to see more websites, more applications, more dashboards, more SaaS products, more landing pages, and more interfaces that look remarkably similar because they were generated from the same patterns, the same component libraries, and the same assumptions about what a modern digital product should look like.

That creates an interesting problem.

When everyone can make something that looks good, looking good is no longer enough.

When everyone can generate a website, I want to be the person who understands what the website actually needs to do. When everyone can generate a dashboard, I want to be the person who knows what information the user actually needs.

When everyone can generate a React application, I want to understand the experience that application should create before worrying about how quickly I can generate the components.

And when AI gives me an answer, I want enough knowledge to question it.

That is the skill I think designers should be investing in.

Not becoming better prompt engineers at the expense of becoming better designers.

Not learning every new AI tool that launches.

Not trying to compete with a machine at producing pixels.

Understanding people more deeply.

Understanding systems more deeply.

Understanding business problems more deeply.

Understanding technology well enough to know what is possible.

And developing enough judgment to know when all of those things are working together and when they are not.

The Opportunity for Designers

I think we are entering a period where actually making things is becoming dramatically easier. That is exciting to me. I want more people to be able to build things. I want designers to be able to prototype without waiting for engineering resources. I want developers to be able to explore design ideas. I want small teams to be able to create products that previously required much larger organizations. But easier production also means we are going to have to become much better at deciding what deserves to be produced.

That is where I think deep UX knowledge matters more than ever. AI has changed how quickly we can create. It has not changed the fundamental reason we create. People still need to understand things. They still need to accomplish tasks. They still get confused. They still make mistakes. They still have different abilities, expectations, contexts, and mental models. Those problems are not going away.

If anything, the tools we have today give us the opportunity to solve them better because we can spend more of our time thinking about the experience instead of simply producing the artifacts.

AI did not make me less interested in UX. It made me realize how important UX actually is. I do not want AI to replace my understanding of UX. I want it to give me more room to use it.