A few weeks ago, I was standing in front of a group of executives in Riyadh, walking them through a design thinking exercise. Empathy mapping. User journey. Problem reframing before solution hunting. The basics. I have been teaching some version of this for more than twenty years — first to product teams in Montreal, then to startup founders in Saigon, now to public sector leaders in the Gulf. The methodology has evolved, the slides have improved, but the core remains the same: before you build anything, understand the human being who will use it.
What struck me that morning was not the content. It was the fact that I was still teaching it at all. After two decades of conferences, bestsellers, certifications, and corporate “innovation labs” plastered with post-it notes, design thinking should be mainstream by now. It should be like project management or financial reporting — an embedded competence, not a workshop. And yet there I was, introducing empathy mapping to a room full of experienced professionals as though it were new.
Then I saw the job data. Design thinking is disappearing from corporate job listings. Not because organizations mastered it. Because they skipped it.
That is the canary in the mine
Here is what the canary is telling us, if we are willing to listen: the human competence layer that design thinking was supposed to build — the capacity for empathy, synthesis, abductive reasoning, the ability to hold ambiguity long enough to let a better question emerge — was never broadly installed. It was performed in workshops and abandoned in workflows. And now, organizations are deploying artificial intelligence into the exact space that competence was supposed to occupy.
The evidence is accumulating from every direction simultaneously, and the convergence is too dense to be coincidental. The research community is publishing findings at a pace I have rarely seen on a single theme: AI does not reduce work, it intensifies it. AI can stifle innovation if left unmanaged. Eighty-eight per cent of organizations have adopted AI in at least one function, but only six per cent see meaningful performance impact. The gap between adoption and absorption is not a technology problem. It is a human infrastructure problem.
What does that mean concretely? It means the chatbot sitting on top of your processes is not making your people faster. In many configurations, it is making them slower — what one researcher calls the “interface cost.” The conversational interface forces human cognition through a bottleneck designed for simple queries, not for the messy, iterative, context-dependent work that innovation actually requires. You would not navigate a ship by typing instructions to a compass. Yet this is precisely how most organizations have deployed AI: a sophisticated engine, accessed through a straw.
I spent two years in Vietnam before arriving in Saudi Arabia, and the experience taught me something that no framework captured at the time. Western innovation methodologies — design thinking included — do not transplant. They metabolize differently in different cultures, not because some cultures are “less ready” but because the human substrate that receives the methodology is shaped by its own history, its own habits of collaboration, its own relationship to authority and ambiguity. What works in a Montreal startup incubator does not work in a Hanoi software house, and what works in Hanoi does not work in Riyadh. The methodology is the seed. The culture is the soil. No amount of seed quality compensates for soil that has not been prepared.
AI is the most powerful seed we have ever produced. And we are planting it in soil that was never tilled.
The organizations that perform — that small, stubborn six per cent — are not better at AI. They are better at being organizations. They invested in the human layer before they invested in the tool. They built the judgement infrastructure that allows people to know when not to use the machine, when to override its recommendation, when to recognise that the pattern it surfaced is an artefact rather than an insight. They treated human capability as the rate-limiting enzyme in the reaction, not as the legacy component to be automated away.
This is not nostalgia for post-it notes. Design thinking in its workshop form deserved to fade — it had become ritual, not rigour. But the competence it pointed toward — the capacity to understand a human being before building something for them — is not optional. It is the substrate without which AI produces confident, well-formatted, entirely wrong answers at scale.
The field is correcting. Researchers are writing about “human magic” and “robot-proof humans” and “genuine human oversight architectures.” These are different vocabularies for the same finding: the constraint has moved. It is no longer the machine’s capability. It is the organization’s capacity to do something intelligent with what the machine produces.
I have spent four years telling my clients in the Kingdom the same thing, in different words: take what I carry and make it yours. Localize it. The methodology is mine. The capability must be theirs. AI does not change that equation. It raises the stakes.
The canary stopped singing a while ago. The question is whether anyone noticed.
How Innovation-Ready Is Your Organization?
Take the free Innovation Maturity Diagnostic — a 3-minute self-assessment based on 30 years of field research across 50+ countries.
Take the Diagnostic →More from Innovation Mastered
