Why Design Thinking Needs an Upgrade for the AI Era
By Nathan Waterhouse ยท 2026-08-21
Almost every organisation I work with has dipped their toe into design thinking. Someone has been on the training, run the sprint, laminated the journey maps; the workshop photos are probably still on a slide somewhere. When I ask what changed afterwards, there is usually a pause. The workshop itself was energising. What it left behind is harder to define.
The framework earned its reputation for good reason. Empathise, define, ideate, prototype, test: it gave a generation of teams a shared language for putting the user back at the centre of decisions that had previously been made in boardrooms with no user in sight. IDEO, where I worked for eleven years, and who did more than anyone to popularise it, had an outsized impact on the world, and nothing I'll say in this article goes against that legacy.
A framework built for a different clock speed
Design thinking assumes three luxuries. It assumes a discrete problem, a season or two to explore it, and a workshop room where the right people can gather. Organisations facing AI-driven disruption now have none of them. The problem keeps redefining itself while you are still empathising with the last version of it, and the tools your team prototyped with in January are a generation behind by June.
Criticism of the framework has been building for a while. Rebecca Ackermann's piece in MIT Technology Review, Design thinking was supposed to fix the world. Where did it go wrong?, traced how a method that promised to democratise innovation ended up, in many organisations, as theatre: sticky notes, a workshop, a report, no change. I think that criticism lands on the wrong target. The failures she describes are mostly failures of containment, organisations treating design thinking as an event, a workshop you run and then file away. That was always the framework's weakest point. A technology that shifts month to month turns it into a structural one: an episodic toolkit cannot keep pace with a moving target.
From workshop to muscle
What replaces it is not a rejection of design thinking's instincts. It is a change of rhythm: from adaptive capacity as a workshop to adaptive capacity as a muscle. Three shifts matter most in practice.
Experimentation loops need to run continuously, in weeks rather than quarters. A team can now build a working prototype with AI tools in an afternoon, which means the constraint has moved. The scarce resource is no longer the ability to make things; it is the ability to decide what to test, and to read the result without flinching. A loop that runs quarterly tests a snapshot. A loop that runs weekly tests a live assumption against a live market.
Decision rights need to sit with the people closest to the work. A central innovation team cannot referee decisions that need to be made faster than they can be escalated. Ronald Heifetz's distinction between technical and adaptive challenges, which I leaned on in From Certainty to Wisdom, is useful here. Rolling out a tool is technical work. Changing how a team senses and responds is adaptive work, and adaptive work cannot be done to people from the centre.
The goal shifts from producing artefacts to building standing capability. Personas, journey maps and prototypes are outputs; the capability that matters is a team's ability to notice a change and respond to it without waiting for permission. In Beyond the Learning Curve I argued that creative thinking is what powers organisational agility. The muscle version of design thinking is that argument made operational: the habits of the workshop, practised until they are simply how the team works.
What survives the upgrade
None of this makes the original toolkit useless. Empathy interviews still surface things dashboards cannot. Prototyping is still cheaper than being wrong at scale. Divergence before convergence still protects a team from solving the first version of the problem it happened to see. What has changed is the container around these habits, not the habits themselves.
The organisations coping best with AI-driven change are not the ones running better workshops. They are the ones that have turned the workshop's habits into a default way of operating: assumptions tested as a matter of routine, users in the room as a matter of course, decisions made close to the work because that is where the information lives.
If your organisation has already done design thinking, sent people on the training, run the sprint, printed the posters, the useful question is not whether to do more of it. The useful question is what comes after it. That is most of the work we now do at Adaptive Edge: helping teams that have outgrown the workshop build the standing capability to keep adapting once the posters come down.