The organisations winning with AI are redesigning work, not deploying more technology

AI is moving faster than the organisations trying to absorb it.

It is already entering workflows, decisions, products and customer interactions, while many organisations are still relying on governance models, data practices and operating structures built for slower forms of change.

The gap is no longer access to capability.

It is the organisation’s ability to learn quickly, redesign work, resolve ownership and make decisions fast enough for AI to create measurable value.

At Digital & AI Edge in June 2026, the discussions showed how AI exposes weak governance, disconnected data, unclear ownership and leadership teams that agree AI matters without agreeing what it should change.

The organisations producing stronger outcomes are changing the work around the technology.

In this article, we’ll examine:

• Why AI is becoming a stress test for organisational readiness
• Why learning speed now matters more than technology access
• Why executive alignment is becoming the constraint on AI value
• How leaders should redesign work before scaling AI further

Reason 1: AI is becoming a stress test for organisational readiness

Most AI programmes still start with the visible layer: the model, the platform, the copilot, the agent or the use case.

Those choices matter, but they do not explain why organisations using similar tools produce very different results.

AI enters an organisation with existing power structures, data habits, governance rhythms, funding models and workflow debt.

Once it is inside the business, those conditions become harder to ignore.

Gabby Fredkin, Head of Analytics & Insights at ADAPT, revealed the gap clearly at Digital & AI Edge.

While 77% of leaders in the room were actively investing in AI agents, only 7% of organisations reported meaningful change to the way work is done.

Investment is consistent, but results vary because organisations differ in how well they absorb technology-led change.

The same tools produce different outcomes.

Reactive organisations buy technology while work stays largely unchanged.

Structured improvers build capability but can spend that capability solving the wrong problem.

Reengineering leaders redesign work in pockets, but without enterprise-wide maturity.

Adaptive organisations combine change capability with first-principles redesign.

Leaders need to know whether AI is landing on work that has been redesigned, data that can be trusted, governance that can scale and executive decisions that can be made at speed.

Dawid Naude, CEO at Pathfindr, argued that many organisations get stuck because AI is delegated too quickly to a central team, CIO, CTO or project office.

AI behaves more like a universal capability than a conventional enterprise system, which means business leaders need to use it directly instead of waiting for technology teams to translate it into strategy for them.

Leaders who do not use the tools are less likely to understand how AI changes judgement, prototyping, customer insight, workflow design or decision-making.

They can approve investment without changing the management system around it.

Mark Pesce, ADAPT Executive Advisor and Co-Founder at Wisely AI, pushed the same point into the economics of work.

He described AI as a repricing of cognition: work that once depended on expensive human thinking can now be performed, supported or accelerated by agents at a different cost structure.

Organisations designed around the old cost of cognition may discover that their processes and business models no longer fit the new economics.

AI shows where thinking is expensive because work is poorly designed.

It exposes unclear decision rights that slow handoffs, gaps in data, authority, or governance that stall teams, and process gains that merely overwhelm the next bottleneck.

A serious AI agenda should begin by pressure testing the organisation:

• Where would AI expose fragmented ownership?
• Where would it increase volume into an already constrained team?
• Where would weak data governance undermine confidence?
• Where would faster execution create risk because accountability is unclear?
• Where would a pilot succeed locally and fail when it meets enterprise reality?

Those questions move the conversation from AI ambition to organisational honesty.

Back to top

Reason 2: Learning speed now matters more than technology access

Powerful models, copilots and agentic tools are increasingly available to every major organisation.

That weakens technology ownership as a source of advantage.

What matters more is how quickly the organisation learns once the technology is inside the business.

Mark Cameron, ADAPT Executive Advisor and CEO & Director at Alyve, described this as the move towards an intelligence-centred enterprise.

Organisations need to stop thinking of themselves as rigid machines built around stable processes, and start thinking as living systems that sense, reason, act and learn.

In that model, the defensible advantage is learning velocity: the speed at which the organisation can absorb information, act on it and improve.

Adoption tells leaders whether people are using a tool.

Learning velocity tells them whether the organisation is getting smarter because of it.

David Walker, former Group Chief Technology Officer at Westpac and DBS and Chair of the AI Council at UNSW, showed why this distinction matters.

His research found more than 70% of the barriers to scaling AI are organisational, including finance systems, governance models, leadership, process change and adoption. Only a much smaller share related to AI technology itself.

DBS did not become an AI leader by beginning with AI.

It first became more nimble: rebuilding technology foundations, strengthening customer focus, enabling experimentation, developing learning habits and redesigning processes.

AI then scaled into an organisation that had already learned how to change.

This challenges annual roadmaps and centrally governed programmes that assume the environment will hold still long enough for the plan to remain valid.

AI is compressing that window.

The stronger organisations are building learning systems, not just delivery systems.

In the Digital & AI Edge panel, Bhaskar Katta, General Manager at Westpac, described this from a practitioner’s view.

Westpac’s journey moved through automation, process and task mining, intelligent document processing and knowledge support before agentic capability entered the picture.

To him, there is no AI without process intelligence.

The organisation has to understand the work before it can improve it.

Otherwise AI automates fragments of a process no one fully sees.

Speaking at the same panel, Jen French, General Manager AI Acceleration at CommBank, described the complementary discipline.

CBA works back from the outcome it wants, using design thinking and systems thinking to reimagine the process before deciding where AI belongs.

Weak AI programmes ask where the tool can be inserted. Stronger programmes ask whether the current workflow should survive.

Andrew Brain, Director, Data AI and Growth at Southern Cross Media Group, made the same point in a disrupted media environment.

His teams are expected to know the commercial value of their work, challenge requests that do not drive outcomes and focus on the business priorities that matter.

In a market where traditional revenue models are being pressured by digital platforms, the value of AI sits in helping the business learn where audience, content and monetisation opportunities are moving.

AI requires a management shift: from project completion to learning cycles, from local productivity to system improvement, from pilots as activity to experiments as evidence, and from adoption metrics to decisions that change because the organisation learned something.

Executives should be measuring the pace of that learning:

• How quickly does a hypothesis become evidence?
• How quickly is a failed experiment stopped?
• How quickly is a successful pattern reused elsewhere?
• How quickly does governance adjust without losing control?
• How quickly does the operating model change when the evidence shows the old work design is wrong?

Advantage is moving to organisations that can learn fast enough to keep rewriting the roadmap.

Back to top

Reason 3: Executive alignment is becoming the constraint on AI value

AI is often framed as a workforce adoption challenge.

Employees need capability, confidence and guidance.

But many of the constraints surfaced at Digital & AI Edge sat above the workforce: executive alignment, governance, funding, decision rights, risk appetite and the willingness to redesign work instead of optimising what already exists.

AI transformation is becoming an executive capability gap.

The issue is visible in the way organisations talk about AI.

Leaders ask which tools employees should use, which pilots should scale, which functions can be made more efficient and which vendors should be selected.

Those questions come too late if the executive team has not agreed what value means, which operating assumptions should change and where AI should alter the way leaders themselves make decisions.

Dawid Naude argued that the leadership team is the common thread in organisations getting material results.

The shift happens when leaders become power users and use AI for strategic work, not just personal productivity.

AI becomes an input into business strategy rather than an output from an isolated AI strategy.

A leader who funds an AI programme but does not understand how AI changes the work is sponsoring deployment, not transformation.

Mark Cameron identified leadership approach as the clearest pattern separating organisations getting significant return from those getting little value.

The top performers are not primarily chasing short-term efficiency.

They connect AI to the organisation’s mission and ask what more the organisation can do that it could not do before.

He also named the executive cohesion problem.

Boards and leadership teams may agree that AI matters, while each function pulls towards its own concern: cost, risk, revenue, product, workforce, customer or compliance.

Everyone supports AI in principle, but the organisation still stalls because the leadership system is not aligned on the change.

If the executive team cannot agree on the value being pursued, the risk being accepted, the work being redesigned and the accountabilities required to scale, AI will expose the fragmentation quickly.

Dan Chesterman, CIO at Teachers Mutual Bank Limited, brought this back to measurement.

He argued that organisations need to baseline performance before applying AI.

Without that discipline, they risk measuring the launch of a tool rather than whether customer experience, process performance or business outcomes improved.

If leaders do not define the before-state, they cannot prove the after-state.

They may count pilots, licences, users or token consumption while missing whether decision quality, customer experience, risk outcomes or cost-to-serve has changed.

Sanjoy Sen, Managing Director and Group Head of Consumer Bank at DBS Bank, showed the opposite pattern in practice.

DBS spent years building cloud-enabled architecture, digital front ends, a common data lake and a platform operating model where business and technology leaders co-create products and share accountability.

AI then became an enabler over a system already designed for data, experimentation and customer relevance.

AI did not make DBS adaptive. Adaptiveness made AI scalable.

Executives cannot ask the workforce to adopt AI while leaving leadership routines unchanged.

Funding needs to support experimentation without losing the discipline of delivery, while governance must move quickly enough to respond to risks that now emerge inside live workflows.

Decision rights also need to move closer to the work, so teams can act on evidence faster, but value still needs to be measured through business outcomes rather than tool deployment.

The executive question therefore needs to move beyond which tools to select and towards how the management system itself must change.

AI should force leaders to ask how decisions, governance, funding and work design must change now that intelligence can be embedded across the organisation.

Back to top

The work has to change before AI can compound

The organisations winning with AI are redesigning how they think, decide, learn and deliver.

That is why Digital & AI Edge pointed to a larger shift across ADAPT’s 2026 agenda.

Scaling AI across the enterprise cannot be separated from executive alignment that unlocks action, or operating model change that sticks.

The three are now linked.

AI scale depends on whether the organisation has the capability to absorb change.

Executive alignment depends on whether leaders can define value, resolve competing priorities and understand the technology deeply enough to make different decisions.

Operating model change depends on whether workflows, governance, data, funding and decision rights can move at the pace AI now makes possible.

When those conditions are missing, AI becomes another layer on top of work that was already struggling.

The practical starting point is an AI stress test:

• Where will AI expose unclear ownership?
• Where will it increase demand into a bottleneck?
• Where will weak data reduce trust?
• Where will governance slow the work or fail to control it?
• Where will executive disagreement stop scale?
• Where will a successful pilot collapse because the operating model cannot absorb it?

These questions determine whether AI becomes advantage or just more activity.

AI reveals organisational capability faster than previous waves of transformation because it moves directly into the work: the decisions, handoffs, data, governance and judgement that shape performance every day.

The fastest organisations are already redesigning those conditions before adding more technology.

Back to top