Summary:
Who this article is for: Designers, brand strategists, and business owners thinking about what AI tools like Lovable actually change about the work of building a website — and what they don’t.
Key takeaways:
- Creative tools have always evolved, but AI tools like Lovable are removing a different kind of friction — the gap between imagining an interaction and building it
- The wall in the creative process hasn’t disappeared. It’s moved: from execution to direction
- A technically flawless website can still be completely ineffective if it lacks clarity, hierarchy, and positioning
- AI search is changing who (or what) arrives at a website first — and a site now has to communicate clearly to both human visitors and AI systems
- Execution is faster than ever, but design judgment matters more, not less
What’s inside:
- How creative tools have evolved — and what feels genuinely different about AI
- Why the struggle in the creative process teaches you things that shortcuts skip
- The shift from building something to knowing what’s worth building
- How AI search is changing the audience a website needs to serve
- What’s worth protecting about web design as execution gets faster
I started doing graphic design about thirteen years ago, and in that time the tools have changed almost beyond recognition.
Some of the things that feel nearly effortless now used to take an unreasonable amount of patience. I remember building GIFs frame by frame in Photoshop. Replacing a background or placing an image into a completely different environment could become its own project. Now Photoshop can do a remarkable amount of that work almost instantly.
Creative tools have always evolved. They’ve always made things faster, more accessible, and considerably easier. What feels different now is the kind of friction they’re removing.
Lovable is a good example. I can describe an interaction I have in my head and, instead of immediately hitting the point where my development skills run out, I can start building it. That’s incredible. There are things I’ve wanted to create for years that lived somewhere between “I know exactly how this should behave” and “I have no idea how to write the code that makes it behave that way.” Vibe coding has started closing that gap, and I love that.
I built this page for you to read alongside this as a proof of concept.
But a small part of me is strangely sad to see some of the friction disappear. Not because I think creative work needs to be unnecessarily difficult, but because the struggle has always been one of my favorite parts.
Where the Process Breaks Down
My creative process rarely begins with a clear solution. It usually starts with an idea, followed pretty quickly by some kind of wall. I try another angle. Then another. Then I come up with something genuinely bad. Then something slightly less bad. Somewhere in that back and forth, the problem starts to look different.
The first idea usually tells me what I thought the problem was. The failed ideas are often what show me what the problem actually is. Maybe I realize the logo needs to work at a much smaller scale than I’d considered. Maybe two pieces of information I thought belonged together are actually competing with each other. Maybe the hierarchy feels wrong because I’m asking someone to care before I’ve given them a reason to.
The mistakes matter. Even the wall matters. Eventually, you come out the other side with more than a better design. You understand what you’re making and why.
AI changes that process because it can get you around the wall incredibly quickly. The question is what we skip when we go around it. Used with intention, AI creates an amazing space for experimentation — I can prototype concepts that might have died before because I didn’t have the development time or technical skill to pursue them. That kind of freedom is exciting. But if the goal becomes avoiding every difficult part of the process, I think something is worth paying attention to. Sometimes the difficult part was doing more than slowing us down. Sometimes it was teaching us what we needed to know.
“The wall isn’t disappearing. It’s moving. It used to sit between an idea and its execution. Now it sits between a tool that can build almost anything and the question of what’s actually worth building.”
Creation Is Getting Faster. Direction Isn’t.
Maybe that’s what’s actually happening. The wall isn’t disappearing. It’s moving.
It used to sit between an idea and its execution. Now it sits between a tool that can build almost anything and the question of what’s actually worth building.
Platforms like Lovable make that shift easy to see. They can generate functional web pages quickly — and in doing so, they make the difference between building something and understanding what should be built impossible to ignore.
A technically flawless website can still be completely ineffective. The buttons can work. The contact form can send. The animations can be smooth and the typography beautifully chosen, and someone can still land there without understanding what the business does, why they should care, or what they’re supposed to do next. Those aren’t engineering problems. They’re problems of clarity, hierarchy, positioning, context, and judgment.
As production gets faster, those decisions matter more. AI is very good at generating answers. It’s a lot less useful when we’re asking the wrong questions.
The User Base Is Shifting Too
For years, UX and UI have centered on a human visitor. We think about how someone scans a page, what captures their attention, where confusion sets in, what builds trust. Those principles aren’t going away. But increasingly, the person may not be the first audience.
Imagine someone asks an AI assistant to find a bookkeeping firm near them that specializes in small construction companies. The assistant reads through a dozen websites and comes back with three recommendations. That person may make a decision without ever seeing the homepage we spent so much time designing.
So what is the AI actually looking for? Plain-language descriptions of services. A clear statement of who the business serves and where. Page structure that makes sense without the visuals. Answers to the questions people actually ask. A homepage that says “We make growth feel effortless” tells a human something about the brand’s tone. It tells an AI almost nothing.
A website now has to communicate clearly to two very different kinds of visitors. A human experiences hierarchy visually — they respond to photography, pacing, interaction, tone, and brand identity. An AI system looks for clear language, structure, context, entities, and relationships. Both need to understand what the company is, what it offers, and why it’s relevant. They just get there in completely different ways. That’s exactly the kind of problem our Build and Grow work is designed to solve.
What I Want to Protect
One of my favorite parts of building a website is creating that first small relationship between a brand and someone discovering it for the first time. For a brief moment, we have their attention. We can invite them to explore. We can help them understand not just what the company does, but what it feels like to work with that company. We can build trust, curiosity, and excitement. That human connection is the part of web design I love most.
I do think we’re moving toward a world where fewer journeys begin directly on the website. AI search will increasingly sit between businesses and the people looking for them. That means we have to be more intentional about both sides of the experience — a website has to be clear enough to be found, interpreted, and surfaced by machines, and when a person does arrive, it should give them something a summary can’t: a sense of the company, a moment of connection, something worth exploring.
Execution is faster than ever. Platforms like Lovable let me build ideas I used to only be able to imagine. Generative tools remove hours of production work. AI search is changing how people find information in the first place.
None of that makes design obsolete. It changes where the design problem lives. The wall is still there. It’s just in a different place now. And the work worth doing is figuring out what a website needs to say, who or what it needs to say it to, and making sure that when a person finally shows up, they have a reason to stay.
Frequently Asked Questions
What is Lovable and how does it relate to web design?
Lovable is an AI-powered platform that lets designers and non-developers build functional, interactive web pages by describing what they want in plain language. It’s part of a broader category sometimes called “vibe coding” — tools that close the gap between having an idea and being able to build it, without deep development skills. It speeds up execution dramatically but doesn’t replace the judgment required to decide what’s worth building.
Does AI make web designers less relevant?
No — it changes where the design problem lives. As execution gets faster, the decisions that require judgment, clarity, positioning, and understanding of the audience become more important, not less. A technically flawless website can still be completely ineffective. The ability to ask the right questions before building anything is the skill that matters more as tools get better at generating answers.
How does AI search change what a website needs to do?
AI search assistants read websites differently than humans do. A human responds to visuals, pacing, photography, and brand tone. An AI looks for clear language, structure, context, and specific answers to common questions. As AI search increasingly sits between businesses and potential customers, websites need to communicate clearly to both audiences — structured and specific enough to be surfaced by machines, and compelling enough that when a human arrives, they have a reason to stay.
Why does creative struggle matter in the design process?
The process of working through failed ideas often reveals what the real problem is — not just what it appeared to be at the start. The first idea usually reflects what you thought you were solving. The iterations show what you’re actually solving. When tools shortcut that process, they can also shortcut the understanding that comes with it. Used with intention, AI creates space for experimentation. Used without it, it can skip the part of the process that was teaching you something important.
What makes a website effective if AI tools can build one quickly?
Effectiveness comes from clarity, hierarchy, positioning, and context — none of which are engineering problems. A website that works technically but doesn’t tell a visitor what the business does, why they should care, or what to do next will underperform no matter how well it’s built. The questions worth asking before building are: What does this need to say? Who (or what) does it need to say it to? And what happens when a real person arrives?
