For years, the digital design industry has been running on a quiet agreement — one that nobody signed and nobody discussed, but everybody honoured. Clients would not ask too many technical questions. Designers would not volunteer too many technical answers. The Figma file would look stunning, the presentation would be confident, the invoice would be premium, and everyone would move on before the product revealed what it was actually built on.
There is a reckoning happening in the IT industry right now, and most people are looking at it from the wrong angle. The headlines say: AI is replacing designers. The LinkedIn posts say: designers must adapt or die. The bootcamp ads say: learn prompt engineering and future-proof your career. All of this misses the one crucial point.
The collapse was always coming. AI made it impossible to hide.
AI is not replacing designers. AI is exposing them — specifically, it is exposing a decade-long pattern that has been running at the centre of the digital industry: the idea that you can design for systems you do not understand, build for platforms you have never learned, and charge premium rates for work that is fundamentally detached from the technical reality it inhabits.
That pattern is now unravelling.
The market is waking up. And the lemon is finally showing through the paint.
The Numbers Are Telling the Story
Before we get into the argument, let us look at what is actually happening in the tech industry.
In 2024, 37% of UX and design organisations experienced layoffs — a figure drawn from the UXPA salary survey, which has tracked the industry for over a decade. Nine percent of individual UX professionals reported being laid off personally. A senior designer reviewing applications for a single UX role received over 300 submissions — a number that tells you everything about the supply-and-demand imbalance that has been quietly building.
Meanwhile, a 2026 study of 500 web design professionals found that 75% report that AI-driven competition has already materially impacted their business in the past year. Entry-level designers at major technology companies now represent just 7% of all hires — a 50% drop from 2019. Design and product teams are consistently among the first to be cut in the ongoing wave of corporate restructuring that shows no signs of stopping.
The industry narrative frames this as an AI problem, an economic problem, a market saturation problem. It is none of those things. It is a competence problem that has been compounding for years, and AI just made it measurable.
The Dirty Secret the Industry Never Wanted to Say Out Loud
For more than fifteen years, the IT and digital product industry quietly filled itself with a category of professional that should not have existed: the digital designer — web designer, UI/UX designer — who refuses to understand digital systems.
Not all designers. Let that be stated clearly. There are extraordinary designers out there — professionals who combine genuine aesthetic intelligence with deep technical literacy, who create work that is both visually exceptional and architecturally sound. Who understand that beauty and function are not opposing forces to be traded against each other, but complementary qualities that a skilled professional delivers simultaneously — precisely because they understand the system well enough to know where creative freedom lives and where technical reality draws the line.
This is not about them.
This is about the other cohort. The one that grew rapidly through the 2010s as design tools became more accessible, as bootcamps proliferated, as Figma removed the last remaining barrier to creating pixel-perfect mockups with no connection to how those mockups would actually be built. The cohort that learned to say the right words — user journey, empathy mapping, design system, component library — without ever understanding the system those words were supposed to serve. The cohort that confused visual polish with professional competence, and found a market willing, for a while, to agree.
Think about any other technical profession and ask yourself whether this would be tolerated.
An interior designer who cannot tell a load-bearing wall from a partition is not just an inconvenience — they are a liability. Structural ignorance of materials, physics, and building codes does not produce a bad-looking room. It collapses the ceiling on someone. The consequence is immediate, visible, and undeniable. The industry enforces knowledge of the underlying system because the cost of ignorance is too immediate to hide.
In IT, the ceiling collapses in slow motion.
The product ships. The prototype is beautiful. The client demo goes smoothly. The invoice is paid. Six months later, the development team cannot add a single feature without breaking three others. The codebase is unmaintainable. The “design system” turned out to be a collection of artboards that shares no logic with the actual implementation. The business is locked into a product that costs more to fix than it cost to build.
But by then, the designer has moved on. And their portfolio screenshot still looks excellent.
The Bridge Nobody Was Building
There is a definition of UI/UX design that the industry has been quietly avoiding because it creates accountability.
The job is not simply to understand users. Any thoughtful, empathetic person can develop user understanding. The job is to connect the system with the users — to serve as the translation layer between technical capability and human need. That is the foundation of design. And that job requires from designers to understand both sides of the bridge they are building. They cannot build a bridge if they only understand one shore.
A designer who does not understand the system they are designing for cannot evaluate whether their design is feasible. They cannot anticipate where their decisions will create technical debt. They cannot have an honest conversation with an engineer about trade-offs. They cannot explain to a client why a feature will take three months instead of three weeks. They cannot catch the moment when their beautiful interaction concept requires a database migration, a new API endpoint, and a backend architectural change.
Design should solve problems, not create them. If design creates problems, then it is not design — it is an art project, or a gamble.
UXPin research puts a concrete number on the cost of this gap: 62% of developers report spending significant time redoing designs due to communication breakdowns — not aesthetic disagreements, but proposals that were technically impossible, or that missed constraints so fundamental they changed the entire approach. This is what AI is now exposing: designers in the IT field who have no working knowledge of the industry they are operating in.
That is not a design problem. That is a competence problem wearing a design problem’s clothes.
What AI Exposed
Here is the irony that the current moment has produced.
AI was supposed to be the great equaliser — the tool that would let anyone produce professional-quality design output without the years of technical learning that used to be the barrier to entry. In one sense, it has done exactly that. Output volume has never been higher. Anyone with a Figma account, an AI subscription, and a few prompt templates can produce fifty screens in a week.
But AI also did something nobody predicted: it attached a dollar figure to incompetence.
A designer who understands their platform uses AI precisely. They know what to ask for, they can evaluate the output, they catch the errors, and they stay in control. AI accelerates their process without replacing their judgment. The cost of their AI usage is modest, absorbed into a workflow that was already efficient.
A designer who does not understand their platform uses AI as a substitute for knowledge. Every vague requirement becomes a long conversation. Every incorrect output requires another round of prompting. Every decision they cannot evaluate independently costs another cycle of back-and-forth. The AI becomes the entire brain of the operation — and brains charged by the token are expensive and so are the mistakes they make.
The result, which anyone working closely with clients in the current market has now witnessed directly, is non-technical designers burning through AI credits at a rate that defies justification. Thousands of dollars in tokens spent to produce work that a technically literate professional could have delivered in half the time at half the price — and with a codebase that will not need to be discarded in eighteen months.
This is not anecdote. The broader pattern is visible in the data on AI token consumption. Inefficiency compounds rapidly. Vague prompts generate more tokens. Incorrect outputs require correction loops that generate more tokens. An inability to evaluate what the AI produced means the conversation cannot be terminated when it should be — it continues, generating more tokens, until the output resembles something the operator cannot distinguish from correct. The price of this is real, and it is landing in client invoices.
The Lemon Market Is Now Open for Business
In economics, a “lemon market” emerges when customers cannot assess the quality of what they are purchasing before they buy it. The seller always knows more than the buyer. And that asymmetry poisons everything.
The digital design market has operated this way for years — with one critical protection keeping it in check: the visual prototype. Because clients could not evaluate the underlying thinking, the technical literacy, or the long-term maintainability of the work, they judged by what they could see. And what they could see always looked good. The Figma file was always polished. The presentation was always confident. The pricing was always premium.
But the product underneath? That was another matter entirely.
Imagine buying a Ferrari for several million euros, only to discover it is running a Ford Escort engine. It looks exactly right from the outside. It does not feel right from the inside. Buy enough of those cars, and you stop trusting any dealer who shows you a Ferrari — regardless of what is actually under the hood. Enough buyers reach that conclusion, and the whole market seizes up. This happens in IT and marketing more than most industries will admit. Businesses stop trusting genuinely skilled professionals because they have been burned so many times by those who only appeared skilled. Eventually, everything starts to sound like a fraud.
Now there are new signals breaking through.
The AI invoice is one of them. The output quality is another. Mentions of “AI slop” — the term for low-quality, generic, AI-generated content — increased ninefold between 2024 and 2025, according to Meltwater data. Negative sentiment around it reached 54% by October 2025. The recognition that AI-generated design has a recognisable aesthetic — Inter font everywhere, purple-to-blue gradients, centred layouts, rounded corners applied without consideration — is growing among clients who have been burned. Integral Ad Science (IAS) analysis found that AI slop sites record conversion rates 91% lower than quality inventory. That is not a number clients will tolerate indefinitely.
And then there is the technical debt audit. The moment a business tries to build on top of what they paid for, and discovers it cannot bear weight — that is the moment the lemon is revealed. It is happening more frequently. It will happen more frequently still as AI becomes the default production tool.
The Figma Problem Nobody Wanted to Address
In May 2025, accessibility expert Adrian Roselli ran automated tests on Figma Sites — the company’s new feature promising to turn Figma designs into live websites. The results were instructive.
One flagship demo site had 210 WCAG accessibility violations. Another had 107. Images had no alt text. Contrast ratios failed basic standards. The HTML output was described as “div soup” — generic container elements rendered in place of semantic HTML, headings, and navigation elements. A technique for drop caps was implemented using background-coloured underscores — a method last relevant in 1974.
This is not a critique of Figma. It is a demonstration of what happens when a design tool is built by and for professionals who have optimised for visual output and abstracted away the technical layer so completely that the technical layer ceases to exist in their workflow. The tools reflect the community. The community reflects the values that were allowed to take root.
When the code that a design tool generates looks like no developer wrote it — because no one who understands how code works was involved in designing the tool’s output — the industry is showing you its foundations. And those foundations matter, because AI cannot fix what it does not understand. Feed a system with incoherent inputs and it will produce incoherent outputs, faster and at greater scale than before.
One of the persistent problems with AI in design workflows is that it overcomplicates decisions in deeply inconsistent ways. If you have not pushed it to its limits on a real production project, that may not be obvious yet. It will be.
AI Slop Is Not the Disease. It Is the Symptom
The mainstream conversation about AI slop treats it as a content moderation problem, a platform problem, a generative AI problem. Clean up the algorithms, label the synthetic content, filter the low-quality output.
This misses the mechanism.
AI slop in the design and digital product space is not produced by AI acting alone.
It is produced by people who use AI as a replacement for knowledge they do not have, to produce work at a scale their actual competence cannot support, at prices the client cannot evaluate critically until it is too late.
The AI did not decide to use Inter font and purple gradients everywhere. The AI learned from what already existed on the web — the accumulated average of everything that came before it — and reproduced it, because the person operating the AI had no knowledge base from which to demand something better. The operator’s ignorance of the system became the AI’s constraint.
The output is mediocre because the direction is mediocre. The direction is mediocre because the person giving it does not know enough to give better direction. A business owner who studied marketing and management can use AI effectively for business strategy and campaign ideation. They cannot use it as a substitute for a senior designer any more than an ethics teacher can substitute for a mathematics teacher. The domain knowledge has to exist somewhere in the room. AI amplifies what is there. It does not supply what is missing.
Technical knowledge is not the enemy of creativity. It is what makes genuine creativity possible. A designer who understands the system knows exactly where the boundaries are — and therefore knows exactly where to push against them. A designer who does not understand the system cannot distinguish a bold creative decision from a technically impossible one. They are not free of the system’s constraints. They are simply uninformed about them.
What Qualified Looks Like
The designers who are not suffering in this market are not the ones who abandoned aesthetics for function, or who traded visual ambition for technical safety. They are the ones who never accepted that trade-off as necessary in the first place.
- They produce work that is genuinely beautiful — because they have the taste, the visual intelligence, and the creative ambition to make it so. And they produce work that is genuinely buildable, maintainable, and effective — because they understand the systems they are designing for well enough to ensure that beauty and function reinforce each other rather than compete.
- They understand component architecture — not because they build it, but because they design with it in mind. They understand performance constraints — not because they optimize the code, but because they do not propose animations that will destroy frame rates on the devices their users actually own. They understand API behavior — not because they write the endpoints, but because they know that a design requiring twelve API calls per page interaction is a design requiring a difficult conversation with the backend team before it goes into production.
- They know where the technical boundaries are, and they work creatively within them — which is what separates real creative problem-solving from the production of pretty pictures that development teams then have to either rebuild or refuse.
- They use AI where it generates genuine value — prototyping, exploration, documentation, repetitive generation tasks. They evaluate AI output critically because they have the knowledge base to recognize what is wrong. Their token consumption is proportionate to their productivity. Their deliverables are buildable. Their handoffs do not require the development team to rebuild the logic from scratch.
The professional who understands the system does not burn through tokens without direction. They use AI where it is needed, delivery time accelerates, quality holds, and the price stays defensible. The product that comes out the other side is visually strong, technically sound, easy to maintain, and built on a foundation that will still make sense two years from now.
That is what the market is now beginning to discriminate toward. Slowly, then all at once.
The Warning That Went Unheeded
This was predictable. For years, professionals who had been in the industry long enough to watch the pattern develop tried to say it plainly: the IT industry has been filling itself with people who are not qualified for the work they are doing, and the bill will eventually come due.
The response was always the same. Design is not engineering. Empathy is a skill. You do not need to code to understand users. The tools have democratised access. Everyone deserves a seat at the table.
These statements are not all wrong in isolation — except one. Design is engineering. It always has been. The two were separated by big tech companies because the industry needed specialists. The industry now needs generalists, and so design is returning to what it has always fundamentally been.
You cannot work in the IT industry — cannot make decisions for IT systems, cannot design for IT platforms, cannot advise clients on IT products — if you have refused to learn how the industry you work in actually works. Not at a code level necessarily. At a systems level. At the level where you understand what you are designing for, what constraints govern your decisions, and what the downstream consequences of your choices will be.
The interior designer does not need to pour the concrete. They need to know what concrete is, what it can support, and what happens when you ask it to do something it cannot do.
Nobody forced this standard on the IT design industry. The industry chose not to enforce it. And for years, the visual sophistication of the tools — and the visual sophistication of the output — was enough to keep the absence of foundational knowledge invisible.
AI has ended that invisibility. It has not done so gently.
What Comes Next
The market correction is underway. It will not be instantaneous — lemon markets correct slowly, because information spreads slowly and customers still cannot perfectly distinguish quality from defect at the point of purchase. There will be more broken applications and websites now, not fewer. Before AI, designers who lacked technical grounding used code generators and website builders to avoid learning the industry they worked in. Those same designers are now using AI to patch the same gap. The shortcut is faster. The ceiling is the same.
The signals of poor-quality products are accumulating. The AI invoices that do not match the value delivered. The products that cannot be maintained. The development teams that have to rebuild what they were handed. The clients who were burned once and are now asking different questions before they sign.
The professionals who will remain are the ones who could have remained before AI arrived — the ones whose value was always real, whose decisions were always grounded, whose work was always built on an actual understanding of the system they were designing for. AI makes them faster. It makes them more productive. It does not make them dependent.
The professionals who will not remain are those who were always, in the precise economic sense of the word, lemons — goods that looked right in the showroom but revealed their defects on the road.
The showroom is closing. The road is where we are now.
If you are a business owner evaluating a design partner, the road is where you need to look. Not at the portfolio screenshots. Not at the mood boards. Ask what their last three projects delivered in measurable business outcomes. Ask how they work with development teams. Ask what they know about the platform they are proposing to design for. Ask them to explain one technical constraint that shaped a design decision they made recently.
The answers will tell you everything the portfolio will not.
The article above describes the problem. The questions below address what working with a studio that has taken that problem seriously actually looks like in practice.
FAQ: Working With Studio Simple
How does studio Simple approach a project differently from a typical design agency?
The most significant difference is that design and development are never separated. Studio Simple is a full-stack design studio — every project is handled by a designer who also builds what they design, with a background in both design and business informatics and over fifteen years of practical industry experience. There is no handoff between a designer who creates and a developer who builds, which means there is no gap where technical feasibility gets lost, misinterpreted, or deprioritised. What gets designed can be built. What gets built matches what was designed. And the decisions made along the way are informed by both sides of that process simultaneously.
The article talks about beautiful design. How does studio Simple make sure design actually performs?
Every project begins with your business — not your brand colours or your aesthetic preferences, but your goals, your customers, and how your digital presence is supposed to contribute to your revenue. Design decisions are evaluated against that context throughout the process, not just at the end. The result is work that is visually strong because quality matters, and commercially effective because that is what you are actually paying for. The two are not in competition. A cluttered interface that fails to guide a user toward a decision is not beautiful regardless of how sophisticated it looks. Clarity, focus, and considered structure in design are not aesthetic choices — they are business decisions with a visual form.
What does “built to last” actually mean for a business owner?
It means you will not find yourself, twelve months from now, unable to update your own website without breaking it — or dependent on an expensive developer to make changes that should take minutes. Studio Simple builds on WordPress with custom modules tailored to your specific content needs, which means your team can manage the site confidently without technical knowledge. It also means the codebase is clean, documented, and straightforward for any future developer to work with if your needs grow. Clients who came to Studio Simple after experiences with poorly built sites — where errors were inherited from previous contractors and the technical debt was significant — consistently describe the contrast in terms of stability, speed, and the ability to finally edit content without anxiety.
How does studio Simple use AI in its work, and does that affect project quality or pricing?
AI is used selectively, as a tool to accelerate specific tasks where it adds genuine value — exploration, prototyping, documentation — not as a replacement for design judgment or technical knowledge. Because the studio has the expertise to evaluate AI output critically, it is not used to substitute for decisions that require understanding the platform, the business context, or the user. This means AI usage is proportionate and purposeful, and its cost is absorbed into a workflow that was already efficient. You are not paying a premium for thousands of tokens spent figuring out what should have been known before the project started.
Studio Simple is based in Croatia — does that affect working with international clients?
Not in practice. The studio works with businesses across Europe — from London and Amsterdam to Stockholm, Zurich, and beyond — and the process is designed for remote collaboration from the first conversation. Communication is clear, structured, and consistent throughout every project, which is one of the qualities clients mention most consistently in their feedback. Time zone differences within Europe are minimal, and the studio’s working language with international clients is English. For businesses based outside Europe — including the US, Canada, Australia, and further afield — the studio will adapt accordingly.
What kind of businesses does studio Simple work best with?
The studio works exclusively B2B, with businesses that take their digital presence seriously as a commercial asset rather than an administrative obligation. The portfolio includes work for yacht charter companies, hospitality brands, legal and engineering firms, and other businesses where brand credibility, digital clarity, and long-term maintainability directly affect client acquisition and retention. The common thread is not the industry — it is the mindset. Studio Simple works best with business owners who understand that a website or brand identity is an investment with measurable returns, not a cost to be minimised, and who want a partner invested in those returns rather than a contractor completing a specification.
What should a business owner do if they have already been burned by a previous design project with an agency that did not deliver?
The first step is an honest audit of what you have — not just visually, but technically. What is the codebase actually like? Can your team maintain the site independently? Are there accessibility violations, performance issues, or structural problems that are limiting your search visibility or user experience? Studio Simple has rebuilt and recovered projects where previous contractors left significant technical debt, and in those cases the audit is where the real scope of the problem — and the realistic path forward — becomes clear. If you have been through a disappointing collaboration and are approaching the next one with caution, that caution is well-founded and worth discussing directly. The right partner will welcome that conversation rather than avoid it.
Reach out and let’s see what we can build together.
