DINO / Design Intelligence Studio
Designing with everything we already know
Great design rarely starts with a blank canvas.
It starts with context.
- 01What are we trying to solve?
- 02What already exists?
- 03What have we learned before?
- 04What patterns should we follow?
- 05What can engineering actually support?
- 06Why were previous decisions made that way?
The requirement is only one input. The rest is the world around it.
- My role
- Framing the Design Brain and Area Brain model, and defining the harness, workflows and prototypes.
- Surfaces
- Mobile and web.
- Period
- 2026 - ongoing.
- Status
- Internal testing ongoing.
TL;DR / Overview
- Problem
- Designers repeatedly reconstruct product context before they can explore or decide.
- My contribution
- Framing the knowledge model and building the harness, workflows and prototypes around it.
- Approach
- Shared product intelligence grounds agent-assisted exploration. Designers retain direction and judgment.
- Knowledge
- Design Brain and Area Brains connect guidance, rationale, product realities and team learning.
- Early estimate
- 2 weeks → 1.5 daysEarly design exploration for a small feature.
- Boundary
- Internal testing. The ambition is a team learning together, not agents replacing designers.
01 / The work around the work
The visible design is only part
of the cost of design.
- TimeUnderstand & explore
- ContextFind & connect
- CollaborationExplain & realign
- WorkMake it tangible
- ReworkRevisit & rebuild
- A time cost
- Days spent understanding the problem, exploring directions and getting to something tangible.
- A context cost
- Finding the right PRD, Figma, implementation, guidance and previous decisions, then reconstructing how they fit together.
- A collaboration cost
- PM explains intent. Design translates. Engineering brings constraints. Design revises. Everyone realigns.
- A work cost
- Mapping flows, recreating familiar patterns, checking states and producing the first thing worth reacting to.
- A rework cost
- Late context reveals something already exists, contradicts a decision, doesn’t fit the product or isn’t feasible.
DINO is designed to change that equation.
02 / From fragmented context to shared intelligence
What would an agent need
to know to design well?
I wasn’t particularly interested in whether AI could generate a screen. It clearly could. I wanted to understand what would make it a meaningful design partner.
A requirement and access to Figma weren’t enough. A designer also carries the product’s language, structure, journeys, principles and the reasoning behind earlier decisions.
A lot of that knowledge exists.
It rarely exists together.
- PRDWhat are we trying to build?
- FigmaWhat have we designed before?
- Design systemWhat components and patterns exist?
- ImplementationWhat exists and works today?
That led to the Design Brain
Not a library of screens.
Not another repository of guidelines.
A layer of intelligence connecting principles, patterns, product structure, terminology, journeys, decisions, rationale, technical realities and exceptions.
Not just what the product looks like.
How the product thinks.
03 / Knowledge has a place
One Design Brain
isn’t enough.
and local decisions
patterns and exceptions
Inherit shared guidance.
Document where and why an area differs.
Each product area develops its own logic: language, mental models, architecture, interactions and history.
I think of an Area Brain as the collective intelligence of that product space. Think about what someone learns after working deeply in an area for a year. That is the knowledge it should begin to capture.
It gives people and agents context to build on, without turning experience into a rigid rulebook.
PrinciplesWhat does good mean in this area?
What do we optimise for? Which behaviours matter most? These beliefs guide decisions when there is no obvious answer, beyond generic design-system guidance.
Terminology & mental modelsWhich concepts must stay distinct?
Names express the product’s mental model, not simply its copy. An agent needs to know what terms mean and which objects must never be conflated.
Information architectureWhere does this screen belong?
Capture primary objects, relationships, hierarchy, global versus contextual actions, entry points and destinations. A reasonable screen still needs to fit the larger product.
Journeys & connectionsWhat happens before and after?
Connect entry points, destinations and branches. Make clear how an action on one surface affects another, rather than judging a screen in isolation.
Patterns & interaction modelsWhat should we follow, or challenge?
Navigation, selection, actions, progressive disclosure, creation, editing and system state all carry established thinking. Knowing the pattern makes it possible to depart from it intentionally.
Decisions & rationaleWhy this way? Does it still apply?
Preserve alternatives and trade-offs: customer behaviour, architecture, accessibility, consistency or performance. A constraint may no longer exist. The outcome alone cannot tell the next designer that.
Constraints, exceptions & lived knowledgeWhat does the team know that isn’t written down?
The technical limitation, unusual edge case, or interaction where two systems meet. This context separates something plausible from something that genuinely belongs in the product.
04 / The team builds the brain
This cannot belong
to one person.
I framed the model and the questions. But the intelligence has to represent the collective understanding of the people building the product.
- DesignPrinciples, patterns,
IA & rationale - PMCustomer problems,
intent & priorities - EngineeringArchitecture,
reality & constraints - ResearchEvidence, behaviours
& unmet needs - ContentTerminology, language
& comprehension
We’re not simply consuming what an AI produces.
We’re collectively teaching the system how our product works.
05 / Put that intelligence to work
Not the answer.
Possibilities.
DINO, the Design Intelligence Studio, is where that intelligence can become useful in the design process. Give it a requirement. The ambition is to understand the world around it before jumping to pixels.
Context
What exists? Which area are we in? What do we know about this problem?
System
Which journeys, patterns, terminology and relationships matter?
History
What was decided before, why, and what have we learned since?
Reality
What is implemented today? Which constraints should shape the exploration?
Illustrative directions, not product screens. Each is checked against what the team knows.
You choose the direction.
- Choose one
- Reject one
- Combine two
- Change the hierarchy
- Challenge the reasoning
- Take another direction
Or reject every option and explore again.
The agent expands the space.
The designer exercises judgment within it.
DINO in action
See the exploration
and the human checkpoint.
Captions could not load. The process descriptions above and summary below remain available.
Options before commitment.
A brief becomes three directions. Choose, iterate, combine or bring another idea before asking DINO to create the Figma.
06 / A different kind of waiting
What happens while
DINO is thinking?
AI takes time.
Gathering context, reasoning across sources, mapping a journey and exploring directions aren’t always immediate. Traditional software gives us a spinner, a progress bar, “Generating…” and asks us to wait.
But if we’re rethinking how people work with agents, I think we should rethink waiting too.
The Campus View recording could not load. The written story remains available.
Waiting time becomes
breathing room.
That’s where Campus View came from. Instead of staring at a loading state, step into the DINO campus.
- Roam and explore.
- Feed the dinosaurs.
- Pet them. Take a breath.
Unwind from all the thinking while DINO does some thinking of its own.
It’s playful, intentionally. Design is cognitively demanding. Sometimes the best thing isn’t to stare at an agent, but to disengage, reset and return with fresh eyes.
Not every second has to become another unit of productivity.
How should work feel when humans and agents operate at different speeds?
Sometimes the agent can work.
And you can pet a dinosaur.
When DINO is ready, come back to the work
with possibilities waiting for you.
07 / Change the conversation
From exchanging information
to making decisions.
Less of the hand-off“Let me explain what
you need to know.”
More of the shared decision“Given what we know,
what should we do?”
- Context
- Gathering enough information to beginStarting from shared understanding
- Collaboration
- Explaining, translating and realigningMaking informed choices together
- Craft
- Producing the first thing worth reacting toEvaluating, combining and pushing ideas further
08 / The economics of design
Speed matters.
Where we spend our time matters more.
For a small feature, in ongoing internal testing.
An early estimate, not a general benchmark or a measured breakdown of every cost.
The bigger opportunity
- Lesstime gathering context
- Lesseffort recreating known work
- Lesscoordination just to get aligned
- Lessrework when constraints arrive late
- Moreexploration before we converge
- Moretime for human judgment
The goal isn’t to squeeze more screens out of a designer.
It’s to change where designers spend their time.
09 / Every project can leave something behind
The intelligence compounds.
Too much design intelligence is temporary. Research teaches us something, engineering uncovers a constraint, and the team makes a trade-off. We preserve the artefact, but often lose the thinking that made it possible.
- PeopleThe team contributes its understanding.
- Shared intelligenceThe Area Brain connects context and rationale.
- DINOAgents bring that intelligence into the work.
- ExplorationMap journeys, surface options and trade-offs.
- Human judgmentChallenge, refine and decide what belongs.
- New intelligenceCapture decisions, exceptions and lessons.
Not just more artefacts.
The decisions we make today can become useful context for tomorrow’s feature. The rationale doesn’t have to stay in a file, a meeting note or someone’s head.
A design organisation that learns together.
People build and question the intelligence. Agents help put it to work. Designers decide when to follow established thinking and when to challenge it.
The next designer doesn’t have to start from zero.
And neither does the next agent.
The future of work I’m interested in
Not simply: “Can AI design faster?”
What if every designer could start with everything the organisation already knows?
People build the intelligence. Agents help put it to work.
Designers bring judgment, imagination and taste.
A better rhythm, not just a faster process.
Between human thinking and machine thinking: sometimes we work together. Sometimes I make the decision. Sometimes DINO does the heavy lifting in the background.
And sometimes I should be allowed to wander around campus and feed a dinosaur while it does.
Spend less time reconstructing what we know.
Spend more time deciding what comes next.
Keep exploring
