For years, we have worshipped at the altar of The Process.
You know the one. User research. Personas. Journey maps. Wireframes. All of it completed before a single line of code is written. All of it delivered in decks that travel up and down the organisation, accumulating feedback, losing specificity, and arriving at the engineering team as a set of documented certainties about a problem that was already different six weeks ago.
The process was not invented by people who wanted to slow things down. It was invented by people who wanted to create accountability in a discipline that had none. It worked — for a while, in a world where the cost of building was high enough that the cost of thinking needed to be proportionally rigorous.
That world is over.
In a recent keynote, Anthropic’s Design Lead Jenny Wen put it plainly: the traditional design process is dead. It has become a bureaucratic loop that prioritises artifacts over end experiences. The deliverable — the wireframe, the persona, the journey map — has replaced the thing the deliverable was supposed to serve. The map has become the territory. And teams are spending the majority of their creative energy building maps.
What is emerging instead she calls the Great Liberation. And I think she is right — though I want to push further than the framing of liberation, because liberation implies escape. What is actually happening is something more demanding than escape. It is a structural shift in what the Product Trio’s job actually is.
From “how” to “what” — the shift nobody has fully named yet
For forty years, digital product design has operated inside what researchers call the WIMP paradigm: Windows, Icons, Menus, Pointers. The designer’s job, in that world, was to make navigation legible. To reduce the cognitive load of moving from A to B. To design the “how” — how a user finds a feature, how a form communicates an error, how a menu reveals its structure.
The new paradigm — sometimes called Architecture 7, or intention-based design — makes a different assumption. The system interprets what you want to achieve, rather than executing clicks in response to deliberate navigation. The user declares an intention. The system acts as an agent. The environment learns from context and memory, and evolves accordingly.
This is not a future state. It is already happening. Claude Cowork can read desktop files, pull Slack threads, and structure a report autonomously from a simple goal — without the user ever opening a menu. The interaction model that would have required a designer to map every screen, every state, and every edge case has been replaced by something that requires the designer to think at a completely different level of abstraction.
The question is no longer: where does the button go?
The question is: how does the system interpret intention?
That is a different discipline. It requires different instincts. And most design organisations have not yet caught up with what it actually demands.
The Wave Model: surfing instead of shipping
One of the most useful frameworks I have encountered for thinking about this shift is what I call the Wave Model of product evolution — and it maps directly to the new mandate of the Product Trio.
Wave 1: Exploration. The Trio uses vibe coding and intuition to test messy, fuzzy ideas in the world. Not in a lab. Not in a slide deck. In real contexts, with real people, building fast enough that the feedback is immediate and specific. This is the phase where the old process would have demanded six weeks of research before a single prototype touched a user. The new approach inverts that sequence — you build to discover, rather than discover to build.
Wave 2: Pilot Systems. Repeatable infrastructure is built to reduce manual coordination. The learning from Wave 1 is stabilised. What worked intuitively gets codified. What failed gets abandoned without ceremony.
Wave 3: Operational Automation. The system becomes algorithmically managed and integrated into the business fabric. The Trio’s role here is not execution — it is to move back upstream, toward the next Wave 1, while Wave 3 runs without them.
The Trio’s mandate is to surf this frontier. The value is not in operating the wave once it is stable. It is in knowing when to move upstream before the wave catches them.
Consider how this played out in a marketplace like Too Good To Go. In Wave 1, the Trio manually tested how bakeries upload surplus inventory. They watched. They adjusted. They built instincts about what friction looked like in practice. In Wave 2, they built automated listing systems that removed that manual coordination. In Wave 3, the system runs without human intervention — and the Trio is already somewhere else, finding the next messy problem worth solving.
The old process would have tried to document the Wave 1 problem into completeness before building the Wave 2 solution. The Wave Model assumes that you can only understand Wave 1 by living in it — and that the sooner you get there, the sooner everything else becomes possible.
Leaving the office — the Flash Build as method
The Great Liberation is not just a change in tools. It is a change in location.
The old process assumed that insight could be gathered in the field and then carried back to the studio, where the real design work would happen. The studio was the legitimate site of creation. The field was where you collected raw material to be processed elsewhere.
The Flash Build inverts this. You build in the field. You prototype in the context where the product will be used. You iterate based on feedback that arrives in minutes, not sprints.
The Nordstrom Flash Build is the example that stays with me. A software team set up inside a flagship store and built a sunglasses app in real time. Within ten minutes of delivering a feature, they had feedback from real customers. Not simulated feedback. Not usability testing with recruited participants behind a mirror. Customers who were actually trying to buy sunglasses, telling them, in the moment, what the tool did and did not do for them.
What they discovered in that environment could not have been discovered any other way. Polarised sunglasses made the iPad screen turn black in portrait mode. That is not the kind of finding that appears in a requirements document or a journey map. It is the kind of finding that only happens when you are in the room where the thing actually occurs.
The Great Liberation requires the Trio to leave the office not as a field trip — but as a structural shift in where the work is done. Real contexts. Immediate feedback. Iteration at a pace that makes the old research cycle feel like a geological timeframe.
This 2011 Nordstrom Innovation Lab video still feels strikingly current—showing a product trio working in-situ with customers to short-circuit the build-test-learn loop—only now, with AI and vibe coding, those same loops can produce real, connected products in minutes.
The Solution-first thinking — and why this is not what it sounds like
Here is the rule that every design education has taught for thirty years: never start with the solution.
I taught it myself. I believed it. And I no longer think it is the most useful framing for the moment we are in.
The old version of this rule was right for a world where solutions were expensive. If building took six months, you needed to be very sure of the problem before you started. The upfront investment in discovery was a hedge against the catastrophic cost of building the wrong thing.
When a working prototype can be vibe-coded in an afternoon, the calculus changes. The cost of starting with a solution hypothesis and testing it immediately is lower than the cost of the research process that was supposed to replace it. The solution becomes a thinking tool rather than a commitment.
But this introduces a different discipline.
Because when solutions become easy to produce, they also become easy to over-polish. It is tempting — especially with today’s tools — to present something that feels finished. Something smooth, responsive, visually resolved. And in doing so, we risk shifting the conversation. The user is no longer reacting to the idea, but to the experience of the artifact itself.
A high-fidelity prototype can create a kind of premature agreement. Not because the problem is well understood, but because the solution is hard to say no to.
So the question is no longer whether we start with the solution. It is how finished the solution is allowed to be when we bring it into the room.
The most useful prototypes today are not the ones that impress. They are the ones that hold back just enough. Clear enough to demonstrate intent. Incomplete enough to invite critique. Functional enough to test behaviour — but not so resolved that they close the conversation.
This is not an argument for building without understanding. It is an argument for a different sequence of understanding — and a different restraint in how we express what we build.
Great intuition — the kind that lets a designer or PM make a reasoned judgment quickly — is not the absence of knowledge. It is the compression of a great deal of knowledge into a pattern-recognition capacity built over years of attention to user behaviour.
The designers and PMs who do this well have spent thousands of hours reading user feedback — Reddit threads, customer support logs, usability session recordings. They have built an internal model of how users think and fail and recover. That model lets them move quickly — and also calibrate how much of a solution to show, and when.
The creation of Claude Artifacts began as a prototype — not a problem statement. A researcher built something. The team looked at it and said: there is something here. The solution revealed a problem worth solving that had not yet been articulated.
But crucially, it was not the polish that made it valuable. It was the signal inside it.
Intuition, trusted and practiced, is the shortest path from deep expertise to working insight. The process was always supposed to be a scaffold for developing that intuition.
The scaffold has become the building.
And now — perhaps — the craft is learning how to build just enough for others to see… without finishing the thinking for them.
What this means for the Trio — and the organisations that deploy them
The shift I have been describing is real, and it is already happening. But it creates a specific challenge for large organisations that have built their operating models around the old sequence.
The Product Trio — PM, designer, engineer — has always been the unit of creation. What has changed is the nature of the creative act. In the old model, the Trio’s job was to follow a manual: do research, generate insights, synthesise them into requirements, hand them to engineering. In the new model, the Trio’s job is to orchestrate intelligence — to work with AI tools, with real-world data, with immediate feedback loops, to find and shape and test ideas at a pace the old process could not have approached.
This changes what good looks like at every level of seniority. Junior members of the Trio need to build instinct faster — because the tools now allow them to test ideas before those instincts are fully formed, which is both a tremendous opportunity and a genuine risk. Senior members need to trust experience over process — and resist the institutional pressure to document everything into completeness before acting.
And organisations need to stop measuring the Trio’s value by the quality of its artifacts.
The wireframe was never the work. The persona was never the work. The journey map was never the work. They were always representations of the work — proxies for the thinking that had been done and the understanding that had been reached.
The work is the understanding. And in the world we are entering, the understanding comes faster, closer to the user, with less translation loss, when the Trio is in the field, building, testing, surfing — rather than in the studio, documenting.
This is the Great Liberation. Not from rigour. Not from craft. Not from the deep responsibility that design work carries when it is done well.
Liberation from the artifacts that replaced the work they were supposed to serve.
It is time to trust the Trio again. Not the process. The people.
This post draws on observations from the Five Mindsets framework developed at McKay Consulting, and on recent conversations at the intersection of AI, product design, and organisational change. The Wave Model maps directly to the sprint arc used in McKay’s Five Mindsets Vibe Coding Sprint — a two-to-three day format for exploring real organisational challenges at the frontier of what AI makes possible.
→ Take the Product Culture Assessment — free, 10 minutes, and a starting point for understanding where your organisation stands today.
→ Explore the Five Mindsets framework and how it applies to the shift described in this article.