One Workflow Is Not a Practice

There's a new kind of dogmatism in design, and it sounds like this: “I just prompt my way through research, generate fifteen variants in Midjourney, test with AI-synthesized feedback, and ship. Why would anyone still do it the old way?”

On the other side of the room, there's the familiar kind: “We need to complete the diverge phase before we converge. Has everyone filled in the empathy map? We can't move to ideation until we've validated our How Might We statements.”

These two people think they have nothing in common. They're the same person wearing different clothes. Both found a process that worked once, decided it was The Process, and stopped thinking.

The AI-native trap

This isn't a piece about how AI is ruining design. It's about what happens when someone mistakes a tool for a methodology.

The new grad who discovered an AI-native workflow has a genuine advantage: they're unencumbered by legacy habits. They see AI as native infrastructure, not an add-on. That perspective is valuable.

But here's what I keep seeing. Someone finds a workflow — Claude for research synthesis, Midjourney for visual exploration, an AI prototyping tool for screens — and because it worked spectacularly on one project, they cement it as their entire approach.

The portfolio becomes five projects that all used the exact same pipeline. The case study is about the workflow, not the problem. “Look how I used AI to do in two days what used to take two weeks.” The process becomes the protagonist.

The moment you hand this person a problem that doesn't fit their pipeline — a complex enterprise flow where the AI can't generate meaningful variants, or a sensitive product where synthesized feedback is dangerously misleading — they freeze. Or worse, they force it through anyway.

One workflow is not a practice. It's a recipe. Recipes break the moment you encounter an ingredient you haven't seen before.

The double-diamond trap

Now the senior. Fifteen years of experience. Can draw the double diamond from muscle memory. Speaks fluently about divergent and convergent thinking, problem space and solution space.

These frameworks describe something real about how good design thinking works. The problem is when the framework becomes the job — followed not because it serves the project but because deviating feels unsafe.

I've watched senior designers spend three weeks on discovery for a problem already well-understood by everyone in the room. The users had been interviewed. The pain points were documented. Engineering had identified the approach. But the process said we weren't allowed to design yet.

That's not rigor. That's religion.

Same rigidity, different aesthetic

Strip away the tools and vocabulary, and both are doing the same thing: optimizing for process comfort instead of problem fit.

The new grad's comfort is efficiency. Their workflow makes them feel fast and modern. Questioning it feels like going backward.

The senior's comfort is thoroughness. Their framework makes them feel rigorous and safe. Questioning it feels like cutting corners.

Neither is asking the only question that matters: does this approach serve this specific problem for these specific users under these specific constraints?

That question is uncomfortable because it doesn't have a reusable answer. It means every project requires genuine assessment instead of pattern-matching to a known playbook.

Flexibility isn't a process. It's the willingness to abandon a process when the situation demands it.

The real skill gap

The industry frames the divide as AI-native versus traditional. That framing is wrong because it assumes the axis is tools.

The actual axis is adaptability.

A new grad who uses AI fluently and knows when to put it down because the problem demands manual depth? Formidable.

A senior with deep methodology who's willing to compress or skip phases when speed matters? Actually worth those fifteen years.

But a new grad with one AI workflow applied to everything is just as limited as a senior with one framework applied to everything. Both are on autopilot. Both produce work that's occasionally brilliant when the problem matches their approach and consistently mediocre when it doesn't.

Process is a tool, not an identity

To the new grad: learn to work without your AI workflow. Take on something domain-specific, politically complex, or emotionally sensitive. That discomfort is where judgment develops.

To the senior: use AI to compress the mechanical parts of your process. Distinguish between phases that create insight and ones you include because they're familiar.

To both: your process isn't your value. Your judgment is. If you can't let go of a process when the situation requires something different, it's not a tool anymore.

It's a crutch. And a crutch, no matter how modern or traditional, still means you're not standing on your own.