Imagine it’s 2030. Product teams no longer use human designers. Business leaders provide their goals, and generative AI models handle the rest. Ideas are produced instantly. Flows are optimized. Wireframes emerge in seconds and are user-tested (synthetically, of course). Then, final designs are implemented directly with code. There are no brainstorms. No team reviews. No ambiguity. Just blazing fast production.
To some, that vision feels like a far-fetched dystopian future. To others, it’s deeply threatening. The question is no longer if AI will affect our work as designers. That part is certain. What we’re scared about is how far it will go and whether it will eventually replace us entirely.
For now (and I believe for the foreseeable future), there’s one critical truth: that great design doesn’t just come from generating and shipping outputs. It comes from understanding how those outputs fit within a real-world context of the human users. And that’s what I want to spend some time exploring.
Context is critical
Several years ago, I had a boss who was known to ask, “What’s the next larger context?” What’s that? It’s the context to the context; which boils down to the context of the problem we’re solving for. It’s the space the problem exists in.
Why does it matter? Because design is more than just making things. It’s about making solutions that fit the shape of the problem. The success of a solution depends not just on how it looks or works in isolation, but how it holds up in the messiness of reality; inside specific systems, for specific users, within specific environments.
That surrounding reality of the problem is what we call “context”. It’s the complexity we inherit when we work on real problems with real people. It includes business goals, technical constraints, organizational dynamics, cultural norms, historical baggage, and user needs that aren’t always clear or well-articulated. And unlike product requirements or best practices, context isn’t static. It shifts over time (sometimes rapidly). It reveals itself slowly (usually methodically). It even contradicts itself.
This is the environment of design. And this is why the work of a designer is as much about sensing and interpreting as it is about producing. Context isn’t something we layer onto a finished design. It’s the target we aim for from the beginning.
Most of design is adapting to context
When you look closely at what designers actually spend time doing, it’s not merely creation. It’s investigating the problem space, exploring how users behave, observing friction, weighing constraints, and testing whether solutions hold up under real-world pressure (typically, with a rough prototype solution in hand).
We revise not merely because of aesthetics, but because it doesn’t yet fit with the circumstances it was meant to serve. We talk with stakeholders not just to gather requirements, but to discover more vector points to map the edges of the problem. We watch how users move through a flow not just to catch usability issues, but to discover what they didn’t say in the interview…more vector points. We’re constantly checking for fit, evaluating whether the solution paths we’re shaping actually make sense in the context it will live in.
This is not extra. It is the work.
Generative AI is great at producing...but that’s only half the job.
To be clear, generative AI brings compelling capabilities to the table. It can produce content, layouts, design variations, and interface copy with impressive speed. It can mimic style, remix components, and generate dozens of solutions in the time it takes us to get oriented to start on one. It’s pretty astounding and is definitely an advantage. Tools that accelerate ideation or offload repetitive work can propel the work of designers and give us more space to focus our brain time on other areas.
But we can’t mistake production (or creation) for design. Generating a range of potential solutions is only half the job. The other half (the more important half) is evaluating which of those solutions best fit the real-world context of the problem and deciding which direction seems most optimal.
And, this is where AI continues to fall short. Because it struggles to see the context of the problem.
AI doesn’t independently understand the world
AI models don’t sense the world directly. (It appears) they don’t have intuition, gut feelings, or lived human experience. They are fed huge amounts of computer language data that’s approximations of what humans have created in that computer language. Then, they generate outputs based on patterns in their training data, creating a confident prediction of what a reasonable next step might look like in response to a given prompt. And while that can create things that appear smart (or even creative), it’s fundamentally a guess. AI doesn’t know the full context of the situation it’s creating for. It doesn’t know the user; it knows some data about the user that it’s been fed. It doesn’t see the politics in the room or the tension in a stakeholder’s voice; it may only know what’s been input into it.
Everything AI processes must be translated and submitted to it. Any AI model is passive and must be fed. It’s not like that’s always on, always listening, always taking in the contextual information around a human. Human behavior is insufficiently converted into structured and unstructured datasets. Emotions become sterile scores. Preferences become textual descriptions. And while the abstractions are useful, they’re also removed from the lived experience they try to represent.
The map is not the terrain. The data is not reality. It’s an inadequate copy.
For now, we design for humans. (Maybe there’s a future where synthetic designs for synthetic?) That means every solution an AI model produces is based on an approximation of what a human might want or need. But approximation is not understanding. And in the messy work of real design, that difference matters more than it seems. Human designers can fall into this trap too. However, the experienced, time-tested, successful designers our world needs are tuned to this understanding and that’s what makes them reliable, successful designers.
Human perception is a design superpower
Humans possess an incredible range of sensing and interpretive capabilities. We see. We hear. We feel. We intuit. We read between the lines. In conversation, we adjust our tone based on how others respond. In collaboration, we notice when something feels off, even if we can’t explain why. These aren’t merely soft skills. They are the foundation of how we work with human complexity.
Designers walk into rooms and feel the dynamic. We observe what people say, and just as importantly, what they don’t. We pay attention to what makes someone pause, what frustrates them, what they smile at. These signals are often subtle, but they shape how we build, how we present, and how we revise. At this stage of technological development, the sheer amount of data this would create can’t be captured and translated into data for AI to process.
It’s this level of perceptive feedback that allows humans to create solutions that are not only functional but appropriate to the exact context we’re solving for. And that’s not something AI can replicate because it’s not something that lives in 1s and 0s alone.
Designers are interpreters of context
The more AI accelerates our ability to produce, the more valuable our human ability to interpret becomes. Designers are not just creators; we’re translators. We sit between the abstract and concrete, interpreting business goals, user behavior, human relationships, and technical constraints into coherent, effective solution paths.
Design requires evaluation and judgment. It requires the ability to navigate ambiguity and change course when the situation shifts. These aren’t extremely difficult inputs to keep feeding into a model.
AI can simulate the output of a design process. But it will struggle to know why one version is better than another in relation to the complexity of the problem context. It can’t sense which direction feels right based on subtleties in context. And until it does, human designers are still required as the interpreters of what fits the problem context.
AI is a tool, not a replacement
So, where does this leave AI? Is it still valuable? Absolutely. AI is already changing how we work. It’s making production faster. It’s unlocking forms of creativity more broadly. It’s exposing and documenting considerations that have been unspoken or missed. It allows us to offload certain tasks. That’s not a threat. It’s an opportunity.
The more we can shift the repetitive and rote parts of our job to AI, the more we can focus on the parts that require a human: the sensing, interpreting, deciding, and guiding. Those aren’t steps in a process. They’re core to the discipline. And they’re not going away.
Until it can feel the shape of the problem…
Remember that imagined future? The one where AI replaces designers entirely? It’s definitely an efficient idea. But it’s missing something essential.
Great design doesn’t happen in theory or in the data. It happens in reality, with real people, under real constraints. It happens in the middle of tension, uncertainty, collaboration, and conflict. It happens when we sit in front of multiple options and evaluate which one fits the complexity of the full view of the problem space. It happens when someone notices something that wasn’t said, or sees a detail that wasn’t obvious. It happens when we interpret the moment (not just the request) and shape our work accordingly. It happens in the nuances.
Until AI can walk into a room, read the energy, sense the silence, gauge the responses, understand the relationships, and have a full view of the real-world it’s creating for…it won’t successfully replace the designers who can.
Because context still matters. And humans live in that context. Humans are the context.