AI can make a design team faster but can it make it smarter?
AI is transforming how designers work in terms of speed, volume and experimentation. The greater opportunity however is to make every project a source of learning, capability growth and collective intelligence that delivers better outcomes for the team and wider organisation.
INSIGHTS
AI
DESIGN
We've been watching the conversation - on how AI is changing design - unfold at other companies in our industry such as Miro and Figma.
There are some valuable
themes emerging, which we wanted to build upon given our vision of elevating 'human intelligence' as the moat for designers.
At Miro’s recent Canvas 26 design leadership panel, nearly half the audience said AI was having its greatest impact on research and synthesis rather than on prototyping or ideation.
That's surprising and revealing. AI isn't simply helping designers produce assets faster, it's entering the thinking process itself: interpreting information, finding patterns, shaping ideas and influencing what teams decide to make. (Something we at wormhole always surfaced and celebrated as a key opportunity for AI.)
Miro calls the broader shift the “Great Inversion”. This is a move from work being 80% doing to 80% thinking, which is an exciting proposition, but it also raises an important question:
If AI affords more time and effort towards thinking, how do we ensure that the thinking is better and more effective?
The speed paradox
The most candid observation from Miro’s panel came from Pundarik Ranchhod of VML:
“We’re definitely able to produce things faster. Whether we’re producing them better, I’m not quite sure yet.”
This is the central tension facing design teams.
AI can accelerate research synthesis, generate concepts and turn ideas into prototypes in dramatically less time. Figma’s State of the Designer 2026 suggests designers are already experiencing those benefits: 89% say AI helps them work faster and 91% believe AI tools improve their designs.
But speed and quality are not the same thing.
When producing a plausible design becomes easier, the difficult work shifts elsewhere.
→ Which problem is worth solving?
→ Which idea is genuinely good?
→ What evidence supports it?
→ What has the team failed to consider?
→ And who is accountable for the outcome?
Execution may well become abundant.
But judgment does not.
This is why expertise, craft and taste become more important in an AI-enabled design practice. A designer is no longer valuable simply because they can produce an interface; their value lies in understanding people, recognising quality and knowing why one direction should progress while another should not.
From the Double Diamond to continuous loops
As the Design Council’s Cat Drew put it:
“Sometimes the prototyping is the research.”
The Miro panel also explored how AI is changing the shape of the design process.
The Double Diamond’s principles of divergence and convergence remain relevant, but the economics of iteration have changed. Instead of progressing through two large phases, teams can run many smaller loops. A hypothesis that once took weeks to investigate might now be explored in hours.
The boundary between discovery and making is dissolving. Teams can build to understand, test to think and learn through rapid experimentation.
This is a significant opportunity, but it contains a hidden risk...
More iterations do not necessarily create more learning. Without continuity, each loop can become another isolated prompt, prototype or discarded direction. Teams generate more possibilities while losing the reasoning that connects them. The finished output remains, but the hypotheses, evidence and decisions behind it disappear.
A faster feedback loop is only valuable if the learning survives the loop.
Design teams therefore need to preserve more than artefacts. They need to retain why an idea emerged, what informed it, what was challenged, what failed and what beneficial learnings should be carried into the next project.
Otherwise, cheaper experimentation produces more activity without building greater intelligence.
Escaping the wall of text
Work is being pulled back into individual chat windows.
Ranchhod on the Miro panel called this the “tyranny of text”. Walls of prompts and outputs that are difficult for others to inspect, challenge or build upon. An individual may work much faster with AI, but the team becomes less aligned.
This matters because design is not just a series of individual tasks. It is a shared process of interpretation. Insight becomes valuable when a team can see it, question it and connect it to the work.
Moving AI output onto a shared visual surface helps but visibility alone isn't enough. Teams also need the context behind the output: its sources, assumptions, decisions and relationship to the broader project.
The goal is not merely shared artefacts, it's shared reasoning.
Without it, AI can create an illusion of progress: more polished work, produced by more people, but all of whom are moving quickly in different directions.
The designer as custodian
The increasingly asked question of the role designers will play, had Anish Joshi, Wormhole's founder highlighting:
“It is how a Designer's innate ability, taste, skill and 'intelligence' work with AI that will differentiate them from a general user using AI for design work."
As AI takes on more execution, designers are moving from makers to stewards. This is not a retreat from making. It is an expansion of responsibility.
Designers increasingly will be responsible for the quality of work created by people and machines. They help define what good looks like, protect the integrity of the process and enable non-designers to participate without losing sight of user needs.
Design systems are an important part of this. They give both people and AI a shared source of components, patterns and standards, but the challenge extends beyond consistency.
A design system can show which component to use. It does not necessarily explain why a decision was made, which evidence informed it, when a method is appropriate or what someone needs to learn to exercise better judgment.
Design teams need living learning systems alongside their design systems to provide the rationale necessary for auditability.
Making every project a learning system
"Being skilled in delegation and guiding isn’t just for creative Directors anymore, it’s for everyone using AI tools”
Will Osborn, Wormhole's Head of Knowledge & Future Skills highlights that Learning Systems will need practitioners to operate differently to be more effective.
A potential learning system for designers could be; research methods appearing during discovery, institutional knowledge surfacing when concepts develop, evidence connected to decisions, and lessons from one project that inform the next.
The purpose isn't to prescribe creativity or replace expertise. It is to help people develop expertise while delivering... turning AI from an answer machine into a growth engine.
This is why we created wormhole as 'Intelligence Infrastructure'. Wormhole integrates credible expert knowledge and institutional information directly into AI guided digital workflows. It helps teams access what they need to know at the moment it becomes relevant, while preserving the context and reasoning surrounding their work - all delivered in a bite-sized digestible manner.
This aligns with the greater opportunity behind Miro’s Great Inversion. Spending less time producing should not be the final goal. The time and intelligence released by AI should help designers think more deeply, make stronger decisions and continually develop their craft.
AI can make a design team faster. But the organisations that succeed will be those that also become better at recognising quality, preserving knowledge, aligning people and growing human capability.
The future of design won't be defined by how much more we can make. It will be defined by how much better we learn while making it.
If you'd like to use AI to design strategically, then grab our field guide: The top 5 things Design leaders are doing differently with AI
Wormhole brings AI-powered guided workflows, integrated contextual bite-sized learning, and artefact generation together so you can 'design the right thing' before you 'design the thing right'.