The Digital Business Ireland business survey Making AI Work for Ireland, published in April, found that 68 per cent of businesses see a lack of skills and training as the single biggest constraint to artificial intelligence (AI) adoption. But what skills are required? AI take-up is happening fast but unevenly, says Cosimo Frontera, research project officer at Eurofound. “Only about one in five European workers use it for work, concentrated among the tertiary-educated [30 per cent of professionals; less than 4 per cent of workers without third-level education].” “In Ireland, enterprise adoption sits at about a fifth of businesses, in line with the EU average, though individual take-up tells a different story: over 30 per cent of Irish workers use AI for work, above the EU average.” Private use is also running ahead of workplace use across Europe: around one in three Europeans have used generative AI in the past three months for work, education or personal use, but only 21.5 per cent used it for work specifically, Frontera adds. Employers want people who use AI to sharpen their judgment, not replace it, says Mark Kelly, founder of AI Ireland. “The goal isn’t cognitive offload. It’s augmentation: better decisions, better questions, higher-quality work. We’re seeing the rise of what Harvard researchers call the ‘thought doer’: employees who bridge strategic thinking and hands-on execution. They encode their own expertise into it and design smarter workflows around it.” The conversation around AI is very dynamic, agrees Annunziata Esposito Amideo, assistant professor in business analytics and management information systems at UCD College of Business. “There are some core foundational elements that must be clear to everyone around ethics and governance; then there is the technical aspect of it. There seems to be a race not to be left behind, and so everyone is trying to get hold of AI tools somehow. However, the most important skill an employer shall look for is ‘understanding and critically analysing the AI output’.” Annunziata Esposito Amideo, assistant professor in business analytics and management information systems, UCD College of Business Employers are increasingly prioritising AI-enabled judgment rather than AI tool proficiency alone, says Bronagh Riordan, AI and data partner at EY Ireland. “While organisations can teach someone how to use a new AI tool, what takes much longer to develop is the ability to ask the right questions, critically evaluate outputs, apply business context and make informed decisions using AI-generated insights. “The most sought-after employees are those who can combine AI with critical thinking, problem-solving, communication skills and industry expertise. The real shift is not about learning another AI tool, it’s about knowing where AI can genuinely help and where human judgment still matters most.” Bronagh Riordan, AI and data partner, EY Ireland AI literacy should be viewed as a baseline skill for everyone, says Laoise Mullane, director, workforce consulting, and AI adoption lead at PwC Ireland: “Understanding how to use AI tools safely, assess outputs critically and apply human judgment in day-to-day work.” Advanced AI expertise is a different, narrower requirement, she says, “needed for specialist roles that develop, govern or deploy AI, requiring deeper skills in areas such as machine learning, data science, AI engineering and responsible AI. The goal is for all employees to be confident AI users, while a smaller group develops the technical expertise needed to build and oversee AI solutions.” Tania Kuklina, director of workforce intelligence and AI at KPMG What KPMG is observing across the market is that, while new roles continue to emerge, the more significant shift is in the evolution of skills within existing roles, says Tania Kuklina, director, KPMG in Ireland. “Demand for AI and data skills is growing rapidly across all core professions, reflecting the increasing integration of AI into everyday work. At the same time, learning by doing remains key. Much of this capability is not built through formal training alone, but through practical application into day-to-day tasks. “Transformations that deliver lasting change share a common characteristic: they are led by all parts of the business.” Successful AI integration is a work organisation problem, rather than just a technical one, says Frontera. “Most enterprise AI pilots show no measurable financial return so far, and what distinguishes the successful roll-out of the AI tools is workflow redesign, manager support and worker involvement.” Rhonda Doyle, country president, Schneider Electric Ireland Skills investment isn’t confined to tech and professional services, says Rhonda Doyle, country president at Schneider Electric Ireland. “Industrial roles are a real growth area, but the pace of competitiveness depends on how confidently we adopt digital and AI-enabled ways of working. Research suggests that, implemented in the right way and with the right skilling, AI could add €250 billion to Ireland’s economy by 2035.” One of the biggest mistakes, she says, is treating AI training as a one-off, or limiting it to a small group of specialists. “More than a third of Ireland’s technology leaders cite a shortage of talent, up 12 per cent from figures last year. Focus needs to be upskilling the workforce and active inclusion.” Cosimo Frontera, research project officer, Eurofound A 2026 EIB study of European firms, says Frontera, “finds that AI raises labour productivity by around 4 per cent, but with complementary investment in workforce training essential to realising the gains”. Eurofound data shows training is the clearest divide between users and non-users: AI users are far more likely to have received employer-paid training (68 per cent against 45 per cent) and to say they’ve learned new things on the job (96 per cent against 73 per cent). Businesses should measure AI training by tracking outcomes, not just participation, says Mullane. “The key is to compare performance before and after training using metrics such as time saved on tasks, output quality, error rates, customer satisfaction, cycle times and revenue or cost impacts. If employees are using AI more but there is no measurable improvement in business performance, then the training may be increasing awareness rather than productivity.” Laoise Mullane, director, workforce consulting, and AI adoption lead, PwC AI-related roles also carry a salary premium, according to PwC’s AI Jobs Barometer, cited by Mullane, with the differential more pronounced in sectors including consumer markets and more moderate in areas such as the public sector. Eurofound’s Frontera notes that PwC’s Global AI Jobs Barometer 2026, based on an analysis of more than a billion job ads, found a “two-track” market “in which ‘professionalised’ roles … are growing at twice the rate and seeing 42 per cent faster salary growth than roles ‘democratised’ by AI, with the wage premium for AI skills reaching 62 per cent.” Entry-level roles most exposed to AI are now seven times more likely to demand traditionally senior skills such as leadership and strategic thinking, compressing the traditional apprenticeship ladder. Lynn Guilbaud, partner in consulting and AI transformation lead, Deloitte In three to five years, we are unlikely to treat “AI jobs” as a separate category, any more than we now talk about “spreadsheet jobs”, says Lynn Guilbaud, partner in consulting and AI transformation lead at Deloitte. “AI fluency will become a baseline skill across knowledge work,” she says. “Customer-facing roles will change significantly, with service, sales and support teams needing to understand how AI shapes customer interactions. Business analysts, project managers and operations teams will also need at least conversational fluency, as will many roles across healthcare and other sectors.”