Most organisations believe AI is a technology challenge. Its quite depressing as it reminds me of the Dot.com boom.
It isn’t.
It is an operating model challenge.
And this is precisely where most AI strategies are quietly failing.
The illusion of progress
Across industries, there is no shortage of AI activity:
- Pilots are being launched
- Tools are being adopted
- Use cases are being explored
On the surface, it looks like progress.
But beneath it, a critical gap is emerging:
There is no coherent operating model to support AI at scale.
AI initiatives are therefore failing to deliver, Gartner and other research groups prove this with published failure rates of 60% to 80% because largely no one thought to operationalise AI and its supporting data management processes and capabilities.
Why the current model doesn’t work
Most organisations are attempting to integrate AI into structures designed for a different era:
- IT governs systems
- Data teams manage pipelines
- Risk and compliance operate in parallel
- Business units drive use cases
This model works for traditional technology.
It does not work for AI.
Why?
Because AI is not static.
It learns, adapts, evolves—and critically, it introduces continuous risk, not just point-in-time risk.
AI breaks traditional boundaries
AI does not sit neatly within one function.
It cuts across:
- Technology
- Data
- Legal
- Risk
- Operations
- Strategy
And yet, in most organisations:
No single operating model connects these domains effectively.
The result?
Fragmentation.
- Models built without governance
- Governance applied without technical understanding
- Risk identified too late
- Accountability unclear
This is not a scaling problem.
It is a structural one.
The missing layer: AI as a managed capability
What is missing is a recognition that AI is not just a tool.
It is a managed, enterprise capability.
And like any enterprise capability, it requires:
- Defined ownership
- Integrated governance
- Standardised processes
- Lifecycle management
- Clear accountability
Without this, AI remains experimental—no matter how advanced the technology appears.
What the future operating model looks like
The organisations that succeed will build an operating model with five defining characteristics:
1. Risk-led design
AI initiatives begin with classification and risk assessment—not just value identification.
2. Embedded governance
Controls are built into the development lifecycle, not applied after deployment.
3. Cross-functional integration
Technology, legal, risk, and business functions operate as a unified system—not in silos.
4. Lifecycle accountability
AI systems are continuously monitored, reviewed, and updated—not treated as static assets.
5. Central coordination with distributed execution
A central AI function defines standards and oversight, while business units execute within that framework.
Why this matters now
The EU AI Act is accelerating this shift.
It introduces:
- Mandatory requirements for high-risk AI systems
- Ongoing monitoring and reporting obligations
- Governance expectations that cut across organisational boundaries
This is not something that can be addressed with policies alone.
It requires structural change.
The real risk organisations face
The greatest risk is not that AI will fail.
It is that organisations will:
- Invest heavily in AI
- Generate early success
- Attempt to scale
…and discover they cannot.
Not because of technology limitations—
…but because their organisation is not designed to support AI responsibly at scale.
Final thought
AI is not just another wave of digital transformation.
It is forcing a redesign of how organisations operate.
Most are still focused on:
What AI can do
Very few are asking:
What must we become to use AI effectively?
That is the real question.
And the organisations that answer it first will define the next phase of competitive advantage.
#AI #AITransformation #DigitalTransformation #OperatingModel #Leadership