Translating AI adoption into business value

When it comes to AI, Irish organisations are investing in pilots and experimentation. While that’s important, those efforts are still not always translating into meaningful business value — because the technology isn’t yet being coupled to the redesign of processes, roles, and operating models. Our 2026 CEO Survey found that only 17% of Irish CEOs reported increased revenue from AI in the last 12 months, while 23% reported lower costs. Most said they had realised neither revenue nor cost benefits.

That gap matters because the first wave of AI has already proved something important: it can help people work faster. Across Irish organisations, AI is accelerating drafting, summarisation, analysis, and routine decision support. Our AI Agent Survey found that 53% of Irish organisations are seeing measurable productivity gains from AI agents, but only 38% say those gains are translating into tangible cost savings.

The question for leaders is no longer whether AI can create pockets of efficiency. It can. The more important question is why those gains so often fail to show up in the outcomes leadership teams care about most: lower costs, faster throughput, better service, stronger conversion, and new revenue.

In most cases, the missing link is not another use case. It’s the business change needed to redesign how work flows, how decisions are made, and where accountability sits. Businesses are still applying AI to workflows built for a pre-AI world. They’re improving tasks but not yet redesigning how value is created across the process. That broader pattern is echoed in MIT Project NANDA’s 2025 State of AI in Business report, which finds that adoption is running ahead of transformation and that many organisations are still struggling to convert pilots into measurable business impact.

Adoption is not transformation

This is the distinction many organisations now need to confront. Adoption means people are using the technology; transformation means processes, roles, decision rights, and performance measures are changing because of it.

Too many AI programmes still sit on top of existing ways of working, rather than being used to redesign the process, the surrounding controls, and the operating model that supports it. A team drafts faster. A function summarises information more quickly. Part of a repetitive task is automated. But the wider workflow often stays intact: the same fragmented systems, the same hand-offs, the same approvals, and the same accountabilities.

That’s one reason the commercial return remains limited. Productivity at task level does not automatically become value at enterprise level. Tactical AI projects can build trust and capability, but on their own they rarely deliver enterprise-wide measurable value. Our 2026 CEO Survey suggests that Irish organisations still lag global peers in the extent to which AI is being applied across multiple business areas, from support functions to strategic decision-making.

This is the trap. If you apply AI to a fragmented process, you often accelerate the symptoms rather than solve the cause. If data is inconsistent, exceptions are poorly managed, roles are unclear, or employees do not trust the output, the technology will struggle to move from experimentation to measurable impact. Our AI Agent Survey makes that point clearly: data issues, mindset, change readiness, and employee adoption remain the biggest barriers to realising value.

The organisations pulling ahead are not simply deploying more AI. They are redesigning business processes enabled by emerging technologies.

Start with end-to-end process redesign

That’s where the conversation now needs to move. The right starting point is not simply, where can we use AI? It’s which end-to-end business processes matter most, and how should they work differently in an AI-enabled model? That shift sounds subtle, but it changes the focus from isolated use cases to end-to-end performance. It forces leaders to concentrate on the workflows where redesign, unconstrained by existing organisational structures, could materially improve cycle time, cost-to-serve, service quality, speed to insight, or customer conversion.

In practice, that means selecting a small number of priority processes and redesigning how those workflows operate. It may involve simplifying hand-offs, clarifying decision rights, updating controls, improving the data needed to run the process, redefining where human judgement remains essential, and reshaping roles around the new workflow. In many cases, it also requires changes to incentives, ownership, and the operating model behind the process.

The organisations making progress are not separating AI deployment from business change. They are coupling the technology to process redesign, workforce adoption and governance from the outset.

That’s where many organisations need to go. Our AI Agent Survey found that 70% of Irish organisations plan to increase their AI-related budgets because of interest in agentic AI in the year ahead. Yet only 9% report broad adoption, and the survey notes that few are using AI agents to fundamentally rethink their operating model and how work gets done.

That’s the missed opportunity. The goal is not simply to make existing work more efficient; it's to rethink how the work should be done in the first place.

What determines whether value scales?

Even well-chosen use cases will stall if the foundations are weak, and trust remains one of the clearest constraints. Our AI Agent Survey found that just 7% of Irish organisations reported high trust in AI agents across multiple functions. Trust was especially low in higher-stakes areas such as financial transactions and autonomous customer interactions. That matters because AI will not be scaled into more material parts of the business unless leaders trust the outputs, understand the controls, and are clear about who remains accountable.

Workforce confidence matters just as much. People are more likely to adopt AI consistently when they understand how it is being used, where oversight sits, and what it means for their role. Our Workforce Hopes and Fears Survey 2025 suggests that Irish workers are more likely to be curious about AI’s impact than worried, but adoption remains uneven and access to learning is not yet universal. While 67% of Irish workers who used AI in the last year said it improved their productivity, only 43% said they had used AI at all in the previous 12 months, and just 10% were using AI tools daily at work. Only 57% say they have access to the learning and development resources they need at work.

The implication is clear. AI value depends on more than the tool. To be successful, organisations need to redesign processes, build workforce confidence, and manage their data responsibly. Across our CEO, AI Agent and Workforce surveys, the pattern is consistent: organisations are starting to see productivity gains, but financial impact will remain limited unless they address the operating conditions that allow AI to scale.  

What leaders should do now

The organisations starting to close the value gap tend to be more disciplined in where and how they act. Here are five ways you can join them:

1. Focus on a small number of high-value end-to-end processes where AI can support measurable business change, not just isolated productivity gains. Avoid spreading effort across a long list of disconnected use cases. Prioritise the workflows where redesign can materially improve cost, speed, quality, or customer outcomes.
2. Redesign the workflow before scaling the technology. Rework not only the task flow, but the controls, roles, decision rights, and ownership that determine whether the process performs differently. Tactical AI projects can build trust and capability, but they are unlikely on their own to deliver enterprise-wide measurable value.
3. Invest in workforce confidence and capability. People need clarity on where AI is used, how outputs are reviewed, what training is available, and how accountability is maintained as ways of working change.
4. Fix data and operating conditions pragmatically. Don’t wait for perfect data or a fully reengineered target state. Prioritise the data, controls, and capability needed for the chosen workflows and improve them iteratively as adoption grows.
5. Measure business outcomes, not just activity. Productivity matters, but leaders also need to track whether AI is improving throughput, service levels, quality, customer outcomes, costs, or revenue. The Irish evidence is now clear: productivity gains are starting to emerge, but financial impact still lags.

AI will not create lasting value as a layer added to yesterday’s operating model. It creates value when organisations couple the technology with redesign of how work flows, how decisions are made, how roles are configured, and how people are supported to act with confidence. That’s the real dividing line.

The organisations that pull ahead won’t be the ones running the most pilots or talking most loudly about AI. They will be the ones disciplined enough to couple AI deployment with process redesign, workforce confidence, and governance strong enough to turn experimentation into measurable value.  

We’re here to help

At PwC, we help organisations turn AI ambition into measurable business value by combining technology expertise with deep sector knowledge and hands-on experience of business transformation. We work with leadership teams to identify the processes that matter most, redesign how work gets done, put the right governance and controls in place, and equip people to adopt AI with confidence. Because we understand the operational, regulatory and commercial realities of different industries, we can help you move beyond pilots and focus on the changes that are most likely to deliver lasting impact. Contact us today to learn more about how AI can reinvent your business.