Oifig IS na hÉireann, the AI Office of Ireland, recently became operational.
The independent statutory body coordinates Ireland’s implementation of the EU AI Act. Its responsibilities include supporting regulatory consistency, facilitating access to technical expertise, and driving AI innovation and adoption.
On one side, regulation is taking shape. On the other, companies and startups are developing new AI products.
But one question keeps coming to mind:
Is it enough to measure success by how many companies use AI?
What is measurable AI business value?
Measurable AI business value is the verifiable contribution an AI investment makes to outcomes such as revenue, costs, productivity, decision speed, customer experience or risk.
The PwC Ireland 2026 CEO Survey makes this distinction visible. Only 17% of Irish CEOs say AI increased revenue over the previous 12 months, while 23% report reduced costs. Most say they have yet to realise either financial benefit.
In PwC’s 2025 AI Agent Survey, 53% of Irish organisations using AI agents report measurable productivity gains, while 38% report tangible cost savings.
These surveys involve different participants and questions, so their results should not be compared directly. Together, however, they suggest something important:
Productivity gains may appear before financial returns.
Why does AI adoption not always create value?
Helping an employee prepare a report faster is valuable. But that alone may not increase company revenue or reduce total costs.
According to PwC Ireland’s analysis, many organisations are adding AI to existing workflows. Tasks become faster, but the gains do not spread across the business unless processes, roles and decision-making are redesigned.
The problem is not always finding another AI use case. Sometimes, the real need is to rethink how work gets done.
What should Ireland’s next stage of adoption look like?
Alongside regulation, the AI Office of Ireland will support AI innovation and adoption. But Ireland’s next stage should focus as much on value creation as adoption rates.
For every AI project, companies could define:
- The starting performance and business outcome being measured
- A revenue, cost, time, quality or customer experience target
- The team or executive accountable for the outcome
- Total costs, including implementation, training, governance and maintenance
- Criteria for continuing or stopping the project
I do not think the question should simply be, “Are we using AI?”
The more important question is:
Is AI genuinely making our business better, and can we measure it?
