The UX Reckoning, Part II: From Interfaces to Intent
- 3 days ago
- 8 min read
We spent 30 years perfecting interfaces, and now they’re disappearing. So, what comes next?
A couple of months ago, I asked a simple question: does AI force us to redesign everything? At the time, it felt like a provocation — a way to challenge assumptions and spark debate. But the more I’ve thought about it, the more it feels incomplete, and potentially even wrong in its conclusion.
What’s actually happening is bigger than a redesign. Through many conversations with operators, designers, and builders, a different pattern is starting to emerge. It’s a fundamental question that goes beyond improving interfaces or modernizing workflows.
We are not redesigning UX. We are replacing it. Not overnight, and not all at once, but in a way that is structural, irreversible, and moving faster than most people inside the industry are prepared to acknowledge.
The Original UX Contract Is Breaking
For the last three-plus decades, UX has operated on a pretty simple contract. The user sits at the center, the interface is the layer they interact with, and the system responds to whatever input they provide. To build better systems, we evaluate interactions and work backwards to design products that people actually want to use. It’s a clean, logical model. And for a long time, it worked.
Everything we built reinforced that structure. Dashboards, menus, workflows, navigation systems—all designed to help users tell software what to do. We spent years refining hierarchy, improving discoverability, and reducing friction so people could move through systems faster and more efficiently.
At its core, this was a model built around instruction. The user issues a command, and the system executes against it. This is the foundation UX has been built on. But it all hinges on one assumption: the user is in control. This assumption is starting to break.
The Interface Is No Longer the Product
There’s a simple way to understand what’s happening here. We like to think technology evolves in clean, linear steps — think of VHS to DVD to streaming. But in reality, the jump we’re making now looks a lot more like going straight from VHS to streaming and skipping the middle step entirely.
The issue is most companies are still designing for that middle step. They’re building what I’d call “DVD UX” — in other words, cleaner interfaces, better workflows, and more polished dashboards. They’re building an agentic layer on top of legacy design. This amounts to an incremental improvement on a model that assumes the interface is still the center of the experience.
But AI-native systems don’t need to play by yesterday’s rules. They don’t depend on users navigating screens or clicking through flows to get things done. In many cases, they don’t even require a traditional interface at all. This is because the value has shifted. It’s no longer about how efficiently a user can move through a system. It’s about how effectively the system can act on behalf of the user. That’s not an upgrade to UX — I’d argue it’s a different paradigm entirely.
From Navigation to Orchestration
Traditional UX was built to solve for navigation. The job was relatively straightforward: helping users find what they need, guide them through a flow, and make the experience feel intuitive. As such, every design decision — layout, hierarchy, interaction — was in service of moving the user through the system with as little friction as possible.
AI systems are solving for something else entirely. They’re not optimizing how users move through software — they’re optimizing how work gets done. At its core, this means interpreting intent, understanding context, taking action across systems — not only within them — and continuously adjusting toward an outcome.
The shift sounds subtle, but it’s not. In the old model, users moved through systems. In the new model, the system moves on behalf of the user. It anticipates what needs to happen, reasons through options, and executes without waiting for step-by-step instruction.
This is where agentic AI actually matters — not as a buzzword, but as a new interaction model. Because once systems can act, the interface is no longer the center of the experience. And once the interface is no longer the center, a lot of what we’ve historically defined as “UX” starts to melt away.
RIP, Pixel-Pushing
For those in the UX trenches, it’s time we be honest about what’s happening. Whether we’d like to admit it or not, a large portion of UX work today is inglorious production design — layouts, components, variants, flows. We're talking pixel precision, endless iterations, and handoffs between teams. It’s not sexy, but it's work, and this agglomeration of tasks makes up the engine room of the discipline. For a long time, this is where most of the billable tasks actually lived, and where most UX designers were employed, especially early in the careers.
The problem is this layer is rapidly disappearing because execution is being automated by AI tools. AI can already generate interfaces, adapt layouts, create variations, connect to real data, and spin up working prototypes in a fraction of the time it used to take.
What once required teams now takes individuals. What used to take weeks now takes days — sometimes hours. The compression is not only real, but also accelerating as AI tools improve. This trend also has a second-order effect people are already feeling but haven’t fully internalizing yet: it doesn’t just change how work gets done — it changes how teams are structured.
What this means is the traditional UX model — especially in agencies and consulting firms — is in deep trouble. As is the case with other professional services professions, UX teams have always lived in a pyramid. Senior leaders at the top, layers of mid-level practitioners in the middle, and junior talent doing the bulk of production work at the base of the pyramid. Like law firms, this model depends on a hefty layer of lower-cost junior folks at the bottom to make financial sense.
But if the production layer is being automated, the pyramid starts to break as its base erodes. What replaces it isn’t just a smaller pyramid. It’s a different shape entirely. Smaller, high-agency teams. More senior-heavy. More cross-functional. Less defined roles between product, design, and engineering — and more shared ownership of outcomes.
Across the industry, we’re already seeing this play out in real time. Fewer designers per project. More overlap across disciplines. Individuals operating at multiples of what used to be possible. This isn’t a temporary efficiency gain. I’d argue it’s a new operating model, which creates a new reality: building is no longer the constraint. Deciding what to build is.
And this is where most teams are not prepared. Because UX has historically operated downstream of product—taking requirements and turning them into interfaces. This model simply doesn’t hold anymore.
Designing Without a Map
There’s an uncomfortable truth at the center of all of this: users don’t know what they want. Let’s face it. they never really did. But in a world of incremental UX improvement, we could rely on research, patterns, and best practices to guide us. We were optimizing within known paradigms, making existing systems easier to use, faster to navigate, more intuitive to understand.
Unfortunately, this playbook doesn’t work anymore. We’re no longer iterating on established models — we’re creating entirely new ones. No one asked for prompt-based interfaces. No one asked for systems that act on their behalf. And yet, that’s exactly where things are going.
This means the design process itself has to change. It means less time spent trying to extract requirements from users who can’t fully articulate what’s possible. It also means less reliance on detailed journey mapping and exhaustive upfront research, and more emphasis on forming strong hypotheses, building quickly, putting real things in front of people, and learning in real time.
We are not researching our way into the future. We are prototyping our way into it. This is also where a second question starts to surface — one that comes up in almost every conversation: If everything becomes a prompt… what actually differentiates anything? If every product has a similar interface… if every system is “AI-powered”… if the front-end starts to flatten… where does the value move? The answer is not where most people are looking.
It’s not in the UI. It’s not in the prompt. It’s in the system behind it. Differentiation shifts to how well a system understands context, how effectively it can take action, how seamlessly it connects across environments, and how reliably it delivers outcomes. It’s about whether the system actually works — not just whether it looks good.
But there’s another layer to this that matters just as much. As execution becomes commoditized, experience becomes emotional. Trust, confidence, clarity, and a sense of control — even when the system is doing the work for you. Those are the things users will remember. Those are the things that will define great products in an AI-driven world.
And importantly, experience can no longer be separated from go-to-market. You can build something technically incredible, but if it doesn’t show up in the right places, with the right narrative, to the right audience, it won’t matter.
Experience is no longer confined to the product itself. It’s the entire system around it — how it’s discovered, how it’s explained, how it’s adopted, and how it evolves over time. This is where differentiation now lives.
The New Role of the Designer
So what happens to UX? It doesn’t go away, but it does change — fundamentally. The designer of the future is not a screen designer. They’re not spending their time perfecting layouts or pushing pixels across interfaces. That work is becoming automated, and with it, the center of gravity for the role is shifting.
What replaces it is a broader, more strategic function. Designers become systems thinkers, product strategists, and behavioral experts. This means they're responsible for shaping how systems behave, not just how they look. They’re orchestrating workflows powered by AI, not just designing the surfaces people interact with.
This means their value is no longer in execution, but rather in judgment. What should this system actually do? When should it act versus wait? What should it prioritize? Where should humans stay involved — and where should they step out? These aren’t interface questions. They’re system design questions. And they require a different skill set. Less focus on tools and production, more focus on framing problems, making decisions, and thinking at a higher level. Less emphasis on outputs, more accountability for outcomes.
This Isn’t a Redesign. It’s a Replacement
If you ask me, the industry is still having the wrong conversation. As we debate the color or shape of prompt boxes, we’re focused on making interfaces better — cleaner UX, more intuitive flows, AI features layered onto existing products. It’s all incremental and assumes the core model still holds, but just needs to be improved.
But that’s not what’s happening. The real shift is that the interface itself is losing its central role. And once this happens, everything built around it — tools, workflows, even entire product categories — starts to move with it.
This isn’t optimization. It’s a change in how humans interact with technology. From clicks to commands, from navigation to intent, and from tools to outcomes. UX isn’t disappearing. But the version of UX we’ve built the last 30 years around the interface-driven model is already starting to fade.
So the question isn’t whether we redesign everything. It’s whether we’re ready to design for a world where the interface is no longer the center of the experience.