Open-ended UX
Designing for people, not progress…oh, and Jane Jacobs ❤️
I used to work in an office that had a tiny kitchen just off the reception area. Deeper than it was wide, it was filled with what you’d expect: a kettle, a sink, a toaster, drawers full of knives (but no forks), chipped mugs, a microwave, and some paper plates for serving birthday cake.
It was fantastic.
That tiny, awkward, frustrating space was the single biggest contributor to company culture. It forced everyone to interact with each other. If you wanted a cup of coffee after a long commute—or a slice of toast after too many beers—you needed to squeeze into that kitchen with everyone else.
It led to friendships, a few weddings, and lots of spontaneous laughter. But I’m certain whoever decided to put that kitchen into such a cramped space, had no idea it would lead to such positive outcomes.
The impact of design is often emergent. People don’t always react the way we imagine they would. They surprise us in how they use the things we build. We can predict it, anticipate what might happen—but we don’t always know. This is why designing for serendipity has been an important part of town planning and architecture for some time [Jane Jacobs wrote the original manifesto on this].
But what about UX? One of the first things we’re taught is to map out a user’s journey. To observe or predict user behavior and define it in a series of steps (often, ideal). Now we’re building products that don’t really do journeys. They’re open-ended. You say what you want—and something magical happens. A simple dialogue where you can ask the AI to do anything. It’s like a magic wand. And I don’t want to storyboard what 3 billion people could do with a magic wand.
The problem with linear thinking
I know user journeys can be useful, but I’ve disliked them for some time. I can’t shake the feeling that they over-simplify. They reduce the complexity of humans into a series of tidy steps, and I think people are inherently messy. They forget why they opened the fridge, splurge when they should save, and have unexpectedly deep thoughts whilst standing in the cereal aisle.
I know some journey maps can be nuanced—layering emotional states, contexts, touchpoints—but they’re still linear. They assume a beginning, middle, and end. And you can see it in the products we build: we design for completeness. Book the flight. Buy the bag. Sign up for the course. We call people to action. We convert. We follow up. We re-engage. Then we start over.
AI disrupts this thinking. A user can take it anywhere, and do pretty much anything. I think that’s why we started calling them co-pilots and agents. The names help us make sense of open-ended experiences by anchoring them to a user’s immediate intent. I’m doing a thing, I need a co-pilot to help me. I can’t be bothered to do a thing, I’ll get the agent to do it.
But what happens when intent isn’t linked to an immediate action or outcome? What if you want to talk about your love life with ChatGPT? Explore whether you’re in the right job with Claude? These aren’t linear tasks. They don’t have a beginning, middle, and end. They meander and they’re all the better for it.
How might we flip, nudge and tilt?
What if instead of trying to turn the messy human experience into an easy to understand journey, we spent our time thinking about how we might shape the messiness? How we might build in pleasant surprises, or tools that help people get the most from the chaos?
More like designing a pinball machine—aware there’s no way to control where the ball goes, but excited by how we might turn that momentum into something special.
In a world of open-ended tools and AI-powered magic, maybe people don’t need help getting from A to B anymore. Particularly if co-pilots are there to show us the way, and agents can do simple tasks for us.
I find this exciting. The idea that we get to build new types of experiences that are less like linear progressions, and more like sandpits or playparks. Jane Jacobs campaigned for this type of thinking in New York City, and it’s what makes it such a vibrant and unique place. Paley Park is my favorite example. A tiny space on E 53rd street. The type of space that might be a parking lot or a low rise building. Some very simple design choices made in the 1960s turned Paley into the perfect pocket park. A huge waterfall dampens the noise of the city. Chairs and tables are free to be moved around according to need or whim. Trees shade people from the sun, and ivy on the walls reminds them it’s a park. But the design of Paley Park is singular. It’s a space that asks people what they’ll use it for. It’s full of potential. The perfect spot for a first date or a break-up. A fun lunch with colleagues or some much-needed time alone.
These spaces—the tiny kitchen, Paley Park—they’re charged with potential because they’re designed with enough room for people to add themselves.
It makes me think the most exciting part of AI is that we can finally stop designing the steps in the journey, and start marking the edge of the dance floor.