UX Caffè
If Anyone Can Create an Interface, What Still Makes a Good Designer?
AI, templates, and design systems are making it easier than ever to produce a credible UI. The designer’s real competitive advantage is shifting from the ability to create to the ability to choose.
Until a few years ago, creating a professional interface required highly specific skills. Designers needed to understand tools, layouts, typography, hierarchy, components, and prototyping. Even reaching a visually credible screen required time, practice, and familiarity with the craft.
Today, that barrier has dropped dramatically. Templates, design systems, component libraries, and especially AI-powered tools allow many more people to create something that looks professional very quickly. We can describe an application in a few lines, generate several visual directions, build a prototype, and even obtain an initial working version of the product. According to Figma’s 2026 AI Report, 70% of product builders say AI already enables them to do work they previously could not do, while 90% believe design is at least as important as it was before AI. Figma
The contradiction is interesting: the easier creation becomes, the more important design seems to become.
The reason may be quite simple.
If everyone can produce, value no longer lies only in production.
It lies in understanding what is actually worth producing.
Making something beautiful is no longer enough
For a long time, strong visual quality was already a differentiator. A clean, coherent, professional UI could clearly separate a polished product from something poorly executed.
Today, the minimum standard has risen.
Component libraries, spacing systems, type scales, UI kits, and generative tools make it relatively easy to produce a pleasant-looking interface. A layout can be perfectly balanced. Colours can be consistent. Buttons can have the right radius. Cards can align properly, and typography can be neatly structured.
But none of this guarantees that the product is good.
A screen can be beautiful and useless.
It can look modern and still be difficult to understand.
It can perfectly follow current visual trends while creating a frustrating experience.
The problem is that when production becomes easy, we risk confusing formal quality with design quality.
They are not the same thing.
Taste matters more when tools become better
In its State of the Designer 2026 report, Figma highlights an interesting figure: 91% of surveyed designers say AI tools are improving their work, while 89% say they are working faster. But the same research reveals something even more important: when tools allow almost anyone to quickly reach a prototype, craft, taste, and the quality of decisions become what really differentiates products. Figma
Taste, however, does not simply mean “having good taste”.
It is not just a personal preference.
It is the ability to understand why one solution works better than another.
It means looking at an interface and recognising that there is too much visual noise even when every individual element is technically correct. It means knowing when to remove something. Understanding when a solution follows a trend but does not fit the product. Recognising when a hierarchy fails to guide the user properly.
Taste is built through experience, observation, and comparison.
It cannot be obtained by pressing a button.
The real problem is no longer generating alternatives
In the past, exploring ten design directions required a significant amount of work.
Today, a generative model can suggest them in minutes.
This fundamentally changes the problem.
The limitation is no longer necessarily the number of ideas we can produce.
The limitation becomes our ability to choose.
If we have ten proposals, which one do we continue developing?
Why?
Which one fits the product?
Which one better addresses user needs?
Which one can be maintained over time?
Which one works with the design system?
Which one risks becoming unnecessarily complex during development?
Figma describes this shift very clearly in its 2026 report: when prototyping becomes fast, deciding what should actually ship becomes even more important. Figma
AI lowers the cost of generation.
It does not automatically lower the cost of bad decisions.
Knowing what to remove becomes a skill
When tools make adding things easy, knowing what to remove becomes even more important.
We can generate another card.
Another chart.
Another feature.
Another step in the flow.
Another variation.
Technically, it is extremely easy.
But every element we add has a cost for the user. It must be seen, interpreted, understood, and remembered.
One sign of a mature designer is precisely the ability to give something up.
Not because it cannot be built.
But because it is unnecessary.
Reducing a flow from five steps to three may be more valuable than perfectly designing all five.
Removing secondary information can improve a screen more than any visual effect.
Simplifying requires judgement.
Adding things is now incredibly easy.
Knowing patterns without becoming trapped by them
Artificial intelligence is extremely effective at reproducing existing patterns. If we ask for a SaaS dashboard, a pricing page, or a financial application, we will probably receive familiar structures.
This is often useful.
Conventions reduce cognitive load and help users understand an interface more quickly.
But a designer needs to know when to use a pattern and when to challenge it.
A sidebar is not automatically the correct solution simply because most dashboards have one.
Three pricing tiers are not mandatory because we see them everywhere.
A bottom navigation does not need five items just because that pattern is common.
A good designer knows conventions.
But does not apply them automatically.
The problem comes first.
Context is not something a prompt automatically knows
One of the main limitations of generative tools is context.
A prompt can describe a product.
But it does not necessarily contain months of stakeholder conversations, research findings, technical constraints, usage data, customer-support issues, business objectives, and real user behaviours.
All of these factors influence design decisions.
Two companies may appear to need the same dashboard and still require completely different solutions.
Two e-commerce platforms may sell similar products but have very different users, logistics, and business goals.
Two applications may share the same feature while needing different flows.
Context is what transforms a generic solution into a specific one.
And it remains one of the areas where designers play a fundamental role.
Asking the right questions
AI is very good at answering.
Designers increasingly need to become good at asking questions.
What is the real problem?
Who are we solving it for?
What happens if we do not build this feature?
What is the most important information on the screen?
What can we remove?
What behaviour do we expect from the user?
How will we measure whether this solution works?
What happens in edge cases?
Which constraints do we need to respect?
These questions are far less spectacular than a screen generated in thirty seconds.
But they often determine the quality of the result much more than the screen itself.
The designer’s role therefore shifts towards problem definition as well.
Before designing a good answer, we need to make sure we have asked the right question.
Visual culture still matters
Tools can produce a lot.
But evaluating what they produce requires a frame of reference.
Visual culture helps us recognise when a solution is generic, when it is following a trend too closely, when a composition is not working, or when an interface lacks personality.
This does not mean studying only digital interfaces.
Editorial design, photography, typography, architecture, cinema, product design, and visual communication can all influence the way we observe and build products.
The more references we have, the more mental tools we have to evaluate what is being proposed.
When AI generates ten solutions, the designer with stronger visual culture is not simply the person who chooses the most beautiful one.
It is the person who can explain why one solution is more coherent and which direction has the potential to become genuinely distinctive.
Design risks becoming identical everywhere
One side effect of the democratisation of tools is homogenisation.
If everyone uses the same libraries, the same templates, the same generative models, and the same prompts, many products will inevitably start to look alike.
The same cards.
The same gradients.
The same dashboards.
The same hero sections.
The same microcopy.
Generative tools work by learning patterns.
This makes them extremely efficient, but it also increases the risk of producing average solutions.
Formally good.
Perfectly acceptable.
And completely forgettable.
The designer’s role therefore also includes introducing intentionality.
Understanding where a convention is useful and where identity needs to be created.
The designer as editor
Perhaps one of the most interesting metaphors for the future of design is the editor.
An editor does not necessarily write every word.
But they need to know what to keep, what to remove, what to rewrite, and which direction the content should take.
Designers may increasingly work in the same way.
AI proposes.
The designer evaluates.
AI generates alternatives.
The designer selects.
AI accelerates production.
The designer maintains coherence, quality, and direction.
In 2026, Nielsen Norman Group described a similar situation as a new phase in which AI enables organisations to build faster than UX can evaluate. The risk is accumulating “UX debt”: products and features generated quickly without being sufficiently understood, tested, or refined. In this environment, UX work increasingly becomes a process of deciding what to keep and what needs to change. Nielsen Norman Group
Speed and quality are not the same thing
AI has made speed one of the main topics in the industry.
We can prototype faster.
Produce faster.
Iterate faster.
But speed is not always the problem that needs to be solved.
If we make the wrong decision in two minutes instead of two hours, we have simply made the mistake faster.
Speed is valuable when it allows us to test earlier, gather feedback, explore alternatives, and improve the product.
Not when it becomes the only measure of process quality.
Designers therefore need to maintain the ability to slow down when necessary.
Analyse.
Question.
Compare.
Test.
Even when the tool in front of us can instantly generate another solution.
A good designer does not compete with the tool
Trying to compete with artificial intelligence on speed would probably be a pointless battle.
A generative tool will be able to produce more variations than any designer.
It can analyse thousands of references.
It can modify a layout in seconds.
The human advantage does not necessarily lie there.
It lies in understanding context.
In sensitivity.
In the ability to manage trade-offs.
In recognising when something is technically correct but humanly wrong.
In defending a decision.
In changing our mind when new evidence appears.
In understanding that sometimes the best solution is not to build anything at all.
These abilities are much less spectacular than an automatic generator.
But they remain the abilities that make the result meaningful.
Conclusions
If anyone can create an interface, the designer’s value does not disappear. It changes position. For a long time, an important part of professional competence was also connected to the ability to physically produce the result: knowing the tools, building layouts, creating components, preparing variations, and turning an idea into something visible. These skills remain important, but they are becoming much more accessible. Artificial intelligence can generate prototypes, templates, and alternatives at speeds that would have seemed impossible only a few years ago, while design systems and libraries make it relatively easy to reach a strong formal standard.
Precisely because of this, the skills that truly make a difference are shifting. Designers need to understand what to choose, what to remove, which problem to address, and which direction to maintain. They need to know conventions without applying them automatically, understand context, interpret user behaviour, and evaluate trade-offs between different needs. They also need enough visual and design culture to recognise when a solution is simply correct and when it actually has a reason to exist.
The point is not to defend the profession from artificial intelligence. That would not be particularly useful. Tools will continue to improve, and many activities that we currently consider a normal part of our work will gradually become automated. That is probably a good thing. If we can spend less time producing ten almost identical variations and more time understanding which variation actually works, we gain more space to do our job better.
But this evolution also brings greater responsibility. When production becomes cheap, it becomes easy to produce too much. When screens look professional after a few minutes, it becomes easy to assume they are already valid. When a prototype works, we may forget to ask whether it is solving the right problem.
And this is where the designer’s value returns.
Not as the person who knows how to use a tool better.
But as the person who knows how to make better decisions.
A UI can be generated.
A product can be prototyped.
A variation can be created in seconds.
The difficult part remains understanding what genuinely deserves to reach users.
The easier AI makes production, the more competitive advantage will depend on judgement.
And perhaps this is precisely what will increasingly distinguish a good designer: not being able to create everything, but knowing what is worth creating.
Follow UX Caffe on instagram
