The next AI productivity leap requires redesigning the organisation
AI is advancing faster than most organisations can turn its capabilities into business impact. The next wave of value will come not from simply adding AI to existing processes, but from redesigning how organisations work around it, writes Wouter Huygen, the CEO of AI and data consultancy Rewire.
AI technology is advancing faster than most companies can absorb it. Increasingly, the constraint is no longer the capability of the models themselves. It is the ability of organisations to turn those capabilities into scalable business impact.
Nvidia’s $12.93 billion acquisition of Hugging Face is a striking illustration of this shift. Nvidia is not simply acquiring another piece of AI technology. Hugging Face has become a critical platform through which millions of developers and more than 200,000 companies discover, evaluate, customise and deploy AI models.
It reflects a broader development across the AI industry. As models become increasingly powerful and accessible, attention is shifting downstream: from building better models to making AI actually work inside organisations.
This is what we call the AI diffusion gap: the widening gap between what AI is technically capable of and what companies are organisationally capable of doing with it.
New technology, old organisations
In one sense, this challenge is familiar. Transformative technologies are almost always initially applied to existing processes and organisational structures.
Consider the electric motor. When factories first adopted electricity, many simply replaced their central steam engine with an electric motor while leaving the rest of the factory largely unchanged. The major productivity gains came later, when factories were redesigned around distributed electric power and individual machines could operate independently.
AI is now entering a similar phase. It should not be seen simply as another technology layer on top of the existing organisation. AI has the potential to become a new operational foundation for how decisions are made, knowledge is applied and work gets done.
That has profound implications for organisational design. Most modern companies were built around human constraints. People have limited memory and context, limited coordination capacity and limited cognitive bandwidth. As a result, work has been divided into departments, management layers, workflows, meetings, approvals, handovers and systems of record.
AI changes some of those underlying assumptions. Intelligent systems can reason across large amounts of information, coordinate activities in real time, learn from feedback and increasingly execute workflows from end to end. Parts of knowledge work – including analysis, coordination, judgement and synthesis – are becoming increasingly automatable at scale.
Yet most organisations are still adding AI to structures that were designed for a fundamentally different world. A smarter chatbot here. A copilot there. Ten individual optimisations inside an unchanged operating model. That may create incremental efficiency, but it misses the larger opportunity.
Don’t just optimise the steps. Redesign the process
An example from our work with a large Dutch telecommunications company illustrates the difference.
The company has a contact centre with around 1,400 employees. When a customer reports a technical issue, an employee traditionally moves through a sequence of steps: identifying the customer, reviewing their history, classifying the problem, searching for information, diagnosing the issue, determining the appropriate action and recording the outcome.
AI can of course make individual steps faster. Speech recognition can transcribe the conversation. A copilot can suggest answers. AI-powered search can help the employee find information more quickly.
But the underlying process remains essentially unchanged.
An AI-native approach starts with a different question: If we were designing this process today, with AI available from the outset, would we design it this way at all?
Instead of optimising a sequence of separate tasks, an intelligent system can bring together customer history, network information and relevant knowledge; diagnose the problem; determine or recommend the appropriate action; execute parts of the solution; and record the outcome.
The employee remains involved where human judgement and interaction add value, while the system takes over much of the information gathering, coordination and execution around them.
The result is not simply a faster version of the existing workflow. It can reduce call duration, unnecessary engineer visits and hardware replacements while improving the customer experience. This is where the larger productivity opportunity lies. Not in making ten steps 20% faster, but in questioning why those ten steps exist in the first place.
From AI adoption to the AI-native organisation
This is the shift towards an AI-native organisation. The term AI-native is increasingly widely used, but the distinction that matters is not whether an organisation uses AI. Almost every organisation does. The real question is whether AI is being added to existing ways of working, or whether processes and operating models are being redesigned around the capabilities of intelligent systems.
Being AI-native does not mean deploying AI everywhere or removing humans from the organisation. It means treating AI as a fundamental design principle: rethinking how knowledge is organised, how decisions are made, how processes are orchestrated and where human judgement adds the most value.
For executives, the important question is therefore changing. It is no longer simply: Where can we use AI? Increasingly, it should be: How would we design this organisation if AI were available from the start?
This does not require a company-wide transformation programme from day one. In fact, the opposite approach is often more effective. Choose one important end-to-end process. Redesign it around AI. Learn from the implementation. Build the required knowledge, data and governance foundations. Then expand.
The transition to an AI-native organisation can start small. But the thinking behind it needs to be radical.
