AI is lifting productivity, but turning those gains into financial results is the next operational excellence challenge. Suzanne Coles, a partner at Curzon Consulting, writing for Consultancy.uk, cites McKinsey's 2026 global survey: 80% of respondents saw improvements in their own productivity, while 37% attributed a positive enterprise-level EBIT impact to AI. In Ireland, William Fry's 2026 Technology Report finds only 11% of larger businesses report a positive return on AI investment to date, though 80% of large firms have invested.
Coles's argument is encouraging, because the gap is a management choice, not a technology failure. The question is why improvements in how people work are not appearing more consistently in business performance. Often, the improvements exist but the decisions needed to realise their value have not been made. Three themes stand out: classifying value, converting capacity, and sharing accountability.
Classifying value comes first. Coles separates delayed value, where credible benefits are still maturing, from stranded value, where proven improvements are blocked by operating structures or management decisions, and unproven value, where evidence is thin. Each needs a different response: milestones, decisions about processes and people, or a fresh look at the original assumptions. Treating all three as a timing problem risks waiting when leaders should be acting.
Converting capacity comes second. Coles has seen process automation cut transaction-processing effort by 20 to 30% while operating costs barely moved, because roles, workflows, and reporting lines were not redesigned. "The process improvement created the opportunity; the business decision created the return." Capacity is also often fragmented: 15 minutes saved daily across 40 people is ten theoretical hours, yet scattered across shifts and roles it may not be usable.
Shared accountability comes third. Finance can set baselines and separate cash savings from avoided costs, while operations knows which processes can change without harming service, so neither can establish the full value alone. Barry Scannell, partner at William Fry, notes that Irish organisations seeing returns have identified practical use cases, invested in skills and governance, and aligned AI with clear commercial objectives: operational efficiency by design.
Three steps follow for Irish leaders. First, name the intended benefit for every AI investment, whether cashable savings, cost avoidance, profitable growth, or strategic value, and never count the same capacity twice. Second, review stranded value now, redesigning roles and workflows where improvements are proven but unrealised. Third, appoint one accountable business leader per investment, an act of operational leadership that links finance and operations.
The lesson from Coles, and from Ireland's own data, is that AI value is made by decisions, not licences. Irish organisations that classify value honestly, redesign work deliberately, and assign clear ownership will turn productivity gains into durable business performance across the economy.



.png)
