AI Is Changing Work

AI adoption is accelerating across organisations. Copilots, intelligent automation, generative AI, and increasingly autonomous agents are being deployed at pace. Yet despite billions in investment, value realisation remains elusive. While most organisations have adopted AI in at least one business function, only a small minority report meaningful outcomes. Recent KPMG research shows that just 8% of organisations have achieved established ROI from AI (KPMG Global AI Pulse Survey Q1 2026) and wider studies suggest fewer than a third are seeing tangible business benefits.

This gap is not driven by a lack of adoption. It exists because organisations are applying AI within existing ways of working rather than redesigning how work gets done. AI is being layered onto operating models, structures, and roles built for a human-only workforce.

The result is an emerging efficiency trap—organisations become incrementally better at executing existing processes but fail to unlock a step-change in performance. The real opportunity with AI is not optimisation. It is reinvention.

Laurène Batkin

Director, People Consulting, Sectors

KPMG in the UK

mail
call

Most organisations are applying AI into existing workflows rather that redesigning how work operates. Real value comes from connecting workflows and redesigning how AI and human operate together – helping turn isolated work into an integrated system that operates more like a smart city.

Laurene Batkin

Director, People Consulting

KPMG in the UK

A Point of View

Re-imagining work to realise the AI opportunity

For the first time, organisations can combine human talent with digital labour to deliver work - not only for routine tasks, but increasingly for analytical, creative, and decision-support activities. This fundamentally challenges how work has historically been designed and managed. For decades, workforce strategy has centred on managing human capacity — the question is no longer just “how many people do we need?”; it is “how should work be designed and distributed across humans, AI, and automation to create value?”

Through our work with organisations navigating this shift, we are seeing a clear pattern emerge: the path to AI value depends less on technology implementation and more on how work, roles and decisions are redesigned. Helping clients address this challenge has revealed five consistent lessons that distinguish organisations generating tangible value from those still struggling to move beyond experimentation.

The data we have now is good for the decisions that we make today, but we need better and different data to support the more complex decisions and to be able to utilise technology (AI) to help use, plan and manage this work.

Megan Butler

Senior Manager AI Workforce

Lesson 1
Lesson 2
Lesson 3
Lesson 4
Lesson 5

Redesign the Work, Not Just the Technology

AI is too often inserted into existing workflows rather than used to redesign them. This limits impact. Task execution is accelerated but work itself does not change. Roles, workflows, and decision-making remain anchored in a pre-AI model.

Organisations must determine:

What work should be automated, augmented, or eliminated
Where human judgement adds value
How work should be restructured to maximise outcomes

Leading organisations are starting to look into the nature of work itself. In one engagement, we deconstructed over 200 roles into 2,000 tasks and 54 activity clusters. This exposed duplication, inefficiency, and inconsistency across the organisation.

By decoupling tasks from roles, we created a clear view of how work gets done. This enabled precise identification of automation and augmentation opportunities—and informed how work should be redesigned. More importantly, it shifted the focus from roles to work.

Tasks and skills are the real units of transformation

Most organisations understand jobs better than they understand work. But AI does not transform jobs - it transforms tasks. Realising value from AI requires moving beyond roles as the primary unit of work design. As work becomes distributed across humans, AI agents, automation, and external capacity, organisations need visibility into how these components interact. Task-level insight provides that foundation.

In practice, this requires linking processes, tasks, and roles. With a global drinks manufacturing organisation, we analysed work at the process level across seven GBS and Finance teams while simultaneously mapping tasks to existing job roles. Our approach surfaced how work flowed across the organisation—highlighting disconnects between roles and processes and revealing where effort was concentrated.

This enabled us to quantify the FTE effort required to deliver end-to-end work and model how automation would shift capacity. It shifted the organisation’s perspective from managing roles to managing work—providing a clear, data-driven view of how activities could be redistributed, where roles could be simplified, and where new capabilities were required to support a more integrated operating model.

Building the data foundation is the hard part

Most organisations underestimate what it takes to enable this shift. A credible AI-enabled workforce strategy depends on a robust data foundation—connecting tasks, skills, roles, workforce supply, and business outcomes. In reality, this data is often fragmented, inconsistent, or incomplete.

Building this foundation requires:

Standardisation
Governance
Integration
Common taxonomies

It is not visible work, but it is critical work.

With a global FMCG organisation, we transformed a static skills repository—developed over two years—into a robust workforce intelligence ecosystem. This involved building a task foundation of over 4,000 tasks across 280 roles and using AI agents to map roles to skills via tasks in near-real time. This made workforce planning actionable, enabling faster and more consistent decisions on capability, investment, and AI opportunity.

Insights do not transform organisations. Decisions Do.

Workforce intelligence has advanced significantly. Organisations can now analyse tasks, model workforce demand, identify emerging skills, and simulate future scenarios. However, insight alone does not create value.

A common failure point is the inability to translate analysis into action. Leading organisations use workforce intelligence to inform strategic choices—reshaping workforce composition, redesigning organisational structures, and prioritising capability investment.

In one case with an FMCG organisation, analysis showed that over 55% of a technology function’s skills would need to evolve to meet future demand. While this insight was significant, the value came where it drove decisions: Reallocation of investment, redesign of roles and prioritisation of capability build. This enabled the organisation to go from reactive planning to proactive, decision-led workforce transformation.

AI transformation is enterprise wide

AI transformation is not a technology programme. It is an enterprise challenge. HR owns workforce strategy. Technology owns deployment. Operations own delivery. Business leaders own outcomes. Each sees part of the system. This fragmentation is the barrier.

AI Transformation demands a new kind of leadership. It requires orchestration: the ability to connect technology, talent, operating and customer capabilities so they reinforce one another rather than compete for attention and investment.

This gap between strategic ambition and organisational capability is becoming one of the defining challenges of enterprise transformation. As change accelerates, competitive advantage increasingly depends not only on knowing what capabilities matter, but on building the organisational capacity to develop, integrate and scale them throughout the organisation.

Leading organisations outperform not because they undertake more transformation, but because they connect technology, operations, workforce and governance efforts into a coordinated enterprise agenda. This creates a more direct path from transformation activity to enterprise performance.

Source: KPMG Transforming the Enterprise 2026

AI transformation succeeds when workforce intelligence becomes the foundation for redesigning work at the task level—orchestrating human and digital labour to unlock value across people, processes, and AI.

Ali Ahsan

Assistant Manager

The Reality

The first phase of AI adoption focused on technology. The next phase is about work.

Organisations that succeed will move beyond roles to tasks, beyond fragmented data to integrated intelligence, and beyond isolated pilots to enterprise-wide transformation. The next frontier of AI transformation is not about deploying more technology, but about redesigning how work is executed, governed and measured across the enterprise. The question is no longer whether organisations can deploy AI, but whether they are prepared to re-architect work, roles and decisions around it to create lasting competitive advantage.