Rethinking the company
AI advantage will depend less on how many tools a company adopts and more on how deeply it redesigns itself around what AI makes possible.
If we built the company today, would we build it the same way?
That is the question I think more leadership teams should be asking.
Access to capable AI is becoming less differentiating as the same models and tools become available to more companies. Giving every employee another AI tool or connecting a few workflows to an API can make repetitive work faster, but it does not fundamentally change the company.
Becoming AI-native is a much bigger organisational question. It means asking how the company should be built around the capability. What people are needed? How should knowledge and data be organised? Where should AI be used? How should decisions be made? How does the business keep learning from customers?
The 2026 Stanford AI Index reports that 88% of surveyed organisations used AI in 2025. At the same time, agent deployment remained in the single digits across nearly all business functions.
AI is only as useful as the foundation it works from. If company knowledge is scattered, data is unreliable and processes are unclear, adding a model does not fix the problem. It can scale it.
Old model versus new model
| Old model | New model |
|---|---|
| AI is a tool used by individual employees. | AI is a capability embedded in how the organisation works. |
| Knowledge lives in people, documents and disconnected systems. | Critical knowledge is organised so people and AI can use and improve it. |
| Good prompts and methods remain personal know-how. | Effective instructions, workflows and learnings become reusable company assets. |
| Automation optimises isolated tasks. | Workflows are redesigned around the best division of work between people and AI. |
| Value is measured mainly as time saved. | Value includes capacity, learning speed, quality, resilience and new business possibilities. |
| Governance is added after the solution. | Access, approval, accountability and data boundaries are designed into the system. |
Individual productivity is not organisational learning
When AI is used primarily by individuals, the productivity gain can be significant. But the knowledge created through that work does not automatically become a shared capability.
One employee discovers a better way to analyse a customer request. Another develops a strong prompt for a recurring task. A third finds a more efficient way to combine information across systems. Unless those learnings are captured and made reusable, the company keeps relearning the same lessons person by person.
This is why I think the shared knowledge foundation matters so much.
The goal is not to document everything. It is to make the company's most important knowledge, instructions, processes and decisions available in a form that can be used across the organisation by both people and AI.
Company knowledge gives AI context. AI can make company knowledge operational.
From assistants to agents
This becomes more important as AI moves from answering questions to performing work.
An assistant can often operate with a short prompt and a document. An agent acting on behalf of a company needs much more: the relevant knowledge, access to systems, process rules, permissions, approval thresholds and a clear understanding of when a person must take over.
The more work technology can perform, the more important the organisation around the technology becomes.
This is also why I do not think the future is simply "more automation". The better question is which work should be automated, which work should be augmented, and where human judgement, accountability, creativity and relationships remain essential.
The European Commission's Industry 5.0 framing is useful because it places human-centricity and resilience alongside productivity. AI should expand the company's capability without making the company more fragile or less accountable.
What companies should build now
- Start with customer and business value. Identify where AI can help the company understand customers better, improve decisions, increase capacity or learn faster.
- Create a shared foundation. Organise the knowledge, data, instructions and workflows that AI needs to operate in company context.
- Redesign the work. Do not automate the old process by default. Reconsider the process around what people and AI are each good at.
- Build governance into the workflow. Define access, data boundaries, approval points and accountability before autonomy expands.
- Turn learning into infrastructure. Capture what works so each experiment improves the organisation rather than only the person who ran it.
The company becomes the product
The point is not that every company becomes a software company. It is that the way the company itself is built becomes part of its competitive advantage.
The companies that create the most value from AI may not be the ones with the most subscriptions or the biggest collection of models. They may be the ones that learn fastest how to redesign themselves around new capabilities.
They learn faster. Important knowledge stays in the organisation. New employees can build on what the company already knows. AI can act with better context. People can focus more attention on judgement, relationships, creativity and the decisions that actually require them.
That is a much deeper transformation than adding a chatbot to the existing company.
The next phase of AI is not only about better technology. It is about building better organisations around the technology.