How organisational design shapes AI scale and value

Most barriers to scaling AI sit inside the organisation rather than the technology stack.

At CIO Edge, David Walker, former Group Chief Technology Officer at Westpac and DBS and ADAPT Advisor, argues that leadership, culture, operating models, finance, and workforce readiness determine whether organisations can turn experimentation into enterprise value.

Key takeaways

• Direct more attention to organisational barriers: Leadership, culture, finance processes, operating models, and workforce change account for far more scaling friction than the technology itself.
• Design for change rather than certainty: Flexible architecture, abstraction layers, and adaptable operating structures allow organisations to change models and infrastructure without wholesale rework.
• Treat nimbleness as an enterprise capability: Organisations that can change direction, absorb disruption, and reinvent quickly are better positioned to capture value from successive technology shifts.

Organisational barriers are holding back AI scale

Technology accounts for only a fraction of the barriers organisations face when moving from experimentation to enterprise deployment.

David analysed 14 major research studies across 105 countries, identifying 367 barriers to scaling AI.

Around 70% were organisational, 20% related to technology and data, and 10% directly to AI technology.

The imbalance matters because investment can concentrate on the smallest category of constraint.

Models and platforms may improve while leadership alignment, culture, workforce engagement, finance processes, and ways of working remain unchanged.

David also points to a deeper mismatch.

Probabilistic technology is being introduced into organisations designed around deterministic systems and processes, creating new demands on governance, decision-making, and tolerance for uncertainty.

Only around 1% of organisations, in the research David cites, have reached a level where they can scale meaningfully.

Closing that gap requires as much attention to how the organisation operates as to the capability being deployed.

Flexible architecture gives organisations room to change direction

Rapid changes in models, infrastructure, and economics make rigid technology decisions harder to sustain.

Trying to predict which model or platform will dominate can lock organisations into assumptions that age quickly.

David recommends using appropriate abstraction layers so components can change as the market and technology evolve.

That flexibility also applies to model selection. Workloads may increasingly use models tuned for specific tasks rather than relying exclusively on large general-purpose models.

The goal is to identify where future change is most likely and avoid hard-wiring those decisions into the architecture.

Models, infrastructure, and supporting components should be replaceable without forcing widespread redesign.

The same principle applies to operating structures. Organisations need enough flexibility to adjust roles, processes, and investment as new capabilities emerge.

Nimbleness turns disruption into an operating advantage

The ability to change quickly shapes how much value organisations can capture from major technology shifts.

David’s work identifies eight characteristics of organisational nimbleness and shows that more nimble organisations respond faster to disruption and achieve materially stronger outcomes.

In the model he presented, the difference equated to at least a threefold return.

DBS provides an example of that progression.

The bank moved from a digital laggard to a recognised digital leader and later generated significant measured value from AI.

That performance came through an organisation capable of sustained reinvention rather than a single technology implementation.

Leaders need to understand how quickly their organisation can change today, identify where friction slows adaptation, and strengthen the systems and behaviours that allow teams to respond.

The organisations best positioned for the next technology shift will be those already built to change when it arrives.