Lorenzo Alvaro is a product designer at Axiom, building scalable visual and UI systems for complex, high-speed trading products.
getty
I’ve watched competitors copy our work. Layouts reconstructed, color systems borrowed, entire screens rebuilt from screenshots. But these copies never feel the same.
This is because users don’t experience products as screenshots. They experience them through what they notice first, where they hesitate, how risk is communicated, how feedback feels, how the product behaves under pressure. That is the part competitors miss, and it’s quietly becoming the most defensible thing in modern software.
AI is making this gap sharper. Templates, component libraries and AI-generated user interfaces (UIs) make polished-looking interfaces cheaper to produce than ever, but visual polish is not product judgment. A screen can look clean and still feel slow to understand, and in decision-heavy products that hesitation lands exactly when users need clarity.
The result is a new kind of product debt: interfaces that look finished but feel generic, inconsistent or thoughtless. Users are starting to call this “AI slop.” In product design, the issue is more specific: It is output without intent.
This changes the basis of competition. If the visible layer is easier to generate or copy, the moat moves beneath the surface. Product feel is what’s left.
Where Product Feel Becomes A Moat
Most of a product’s feel is felt before users can name it. They can’t explain why one product feels clear and another noisy, but they sense it in how each behaves. That sensing compounds.
Products that retain users are not always the ones with the most features or the cleanest visuals. Often, they are the ones that feel most natural to use in context. Apple’s iOS is a clear example; its advantage shows up in daily use, coming less from having every feature than from restraint, cohesion and dependability across thousands of small interactions.
In a trading interface, the difference between a confident execution and a missed entry is often half a second of hesitation on a confirmation screen. That hesitation is rarely a logic problem. It is a feel problem. Users will not say the hierarchy is off. They will just feel slightly less in control, and act on it.
This is where taste stops being about aesthetics. Taste is the ability to make the right judgment about how a product should feel in use. Competitors can copy the surface. They almost never copy the judgment behind it.
Why AI Raises The Value Of Judgment
AI changes what becomes scarce. When more teams can produce polished-looking interfaces, the advantage moves from producing more output to knowing which output deserves to exist.
AI can generate a clean flow, but it does not understand where the product should create friction, remove doubt or make an action feel deliberate. These are judgment problems, and judgment is what you bring to AI, not what AI brings to you.
My engineers have found AI to be a real speed unlock. But the output almost never lands cohesively on the first pass. Therefore, make sure to treat AI-generated UIs as a starting point, not a finished product. Otherwise, the hierarchy is flat when one value should dominate.
For example: There may be too much information shown when half of it should be cut so the rest lands harder. Contrast between adjacent elements is muted, so nothing leads the eye. AI often can get you 70% there, but someone has to make the decisions it cannot: what to remove so a trader parses the screen in under a second, which value earns the most weight, where the eye should land first. That filtering layer is the judgment, and it does not scale by adding more AI.
In short: AI increases what can be made. Judgment decides what should be made, what should be removed and how the product should feel when it reaches the user.
What Leaders Should Look For
If not clear by now, the takeaway is not to avoid AI, but to know where to use it and where to push back. I find that AI is most valuable when it accelerates exploration. This means using it to generate options, explore directions and reduce the cost of trying things.
The risk is using AI where the product needs judgment, not just output. Be especially careful around foundational stages and moments that shape trust: primitives, design language, onboarding, navigation, confirmation states, pricing and error handling. Any flow where the user has to make an important decision is where product feel shows up.
In those moments, the better prompt is not: “Can you make the history data show in this page high-end polished design and intuitive.”
It is, for example:
• “Make the historical data table, filtering controls and date selector cohesive with the rest of the design language using @designsystem.md, referencing the existing dashboard table and analytics card components for structure and implementation patterns.”
• “Show the best hierarchy and order of information for optimal parsing for this page context and users.”
• “In portfolio overview, cut elements for faster parsing while keeping the existing performance chart, position summary and risk indicators.”
The value is in the granularity. Prompts should pressure-test judgment and intent, not just produce raw output.
Conclusion
The traditional workflow loses intent at every handoff. A designer makes a decision in Figma, hands it to engineering and something gets lost. More gets lost when implementation hits an unanticipated constraint. By the time the feature ships, the original feel has been diluted by a dozen small compromises.
When AI is integrated as a system across design and engineering, that loss collapses. The same judgment holds from initial decision through shipped code. Product feel survives.
Start using AI as a system, not as a tool. The gap will not look dramatic at first, but it will show up in how the products feels.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
