The hype surrounding all things AI following the launch of ChatGPT in late 2022 has abated somewhat and organisations and commentators are taking a more realistic view of the technology. While most surveys indicate that a majority of businesses intend to adopt it in the coming years, many organisations are still struggling to identify practical use cases for it.
The pace of AI adoption is relatively slow, and doesn’t match the high expectations for the technology, according to David Lee, chief technology officer with PwC Ireland.
“PwC’s 2025 GenAI Business Leaders survey showed that 86 per cent of Irish business leaders believe that the overall impact of AI on Ireland’s economy in five years’ time will be positive; while 82 per cent believe that AI will deliver increased efficiency in their employees’ time at work. However, scale adoption is at a slow pace – while two thirds (67 per cent) are either at the testing or partial implementation stages of AI adoption, only 6 per cent of respondents in the same survey reported widespread adoption of AI.

“The gap we are seeing here is reflective of the maturity of adoption in Irish organisations, and as more organisations progress through their journeys from testing AI through to wider-scale adoption, we expect to see this gap decrease.”
READ MORE
This gap between enthusiasm and application is quite natural and has been witnessed in previous waves of technology, according to Liam McKenna, partner in the consulting practice in Forvis Mazars, who points out that the time taken to bridge the gap is reducing.
“It took a long time from the development of the PC to when the spreadsheet emerged as the killer app,” he notes. “And it took 10 to 15 years for businesses to make money on the internet for the first time. The cycle was much shorter after the introduction of the smartphone. The cycle is similar, but the time span is reduced. We expect that within the next three to five years we will see fundamental change in business and organisations of all types as a result of AI.”
McKenna refers to the Gartner definition of the “hype cycle”, which is composed of five phases: the technology trigger, the peak of inflated expectations, the trough of disillusionment, the slope of enlightenment, and the plateau of productivity.
“We are going through it now with AI, but it will be much shallower. There are practical use cases at the moment. For example, we are working quite a lot with hospitals where it is being used to enable doctors to spend more time with patients. We are also seeing organisations with really clear and well understood business cases which have been able to roll out the technology successfully.”
The majority of organisations are still at the innovation and exploration phase of their AI journeys, adds Martin Duffy, head of GenAI with PwC Ireland. “Business leaders are approaching AI adoption in a considered manner,” he says.
“Businesses have worked hard to establish relationships of trust with their staff and customers, and they want to ensure that these are sustained on their AI journey. So, they are cautiously experimenting with the technology. They are learning from their innovation activity that the safe and successful deployment and sustained adoption of AI is a complex process that requires planning and co-ordination across the organisation.”
There are variations in AI adoption depending on geographical location, firm size and by economic activity, says Erik O’Donovan, head of digital economy policy with Ibec.
“The EU lags the USA in relation to AI adoption and there are geographical divides in adoption across the EU itself. AI is beginning to deliver financial impact across business functions, but most companies are early in their journeys. In Ireland, AI adoption in SMEs and the public sector tends to lag [behind] large enterprise. AI use is currently more intensive in the ICT sector, professional, scientific and technical service activities.”
Economic enthusiasm for AI is based on productivity and trade opportunities, he adds. “However, that means enabling conditions where opportunities can be understood and realised are not a given. Regulatory uncertainty; lack of capacity, including skills; and access to capital are challenges.”
Prof Andreas Hoepner of the UCD School of Business isn’t sure that the gap is as big as many people think. He points out that there is a difference between using AI in areas such as marketing or customer support and deploying it for new product development or operational efficiencies.
“Maybe they don’t want to talk about that until they are sure they are ahead of the competition,” he says. “It is similar to IP [intellectual property]. If it is good IP, they want to keep it secret for as long as possible. There may be a lot more going on beneath the surface than we know about.”
On much the same note, David O’Sullivan, director, consulting, with Forvis Mazars, points out that the number of active users of GenAI is increasing and that there is still growth in shadow use, where people in organisations are using it without necessarily having authorisation to do so.
“A report by Trinity College Dublin and Microsoft showed a lot of use without policies being in place in organisations,” he says. “Organisation-wide adoption hasn’t reached the level people thought would happen, though.”
He believes this is likely to change quite quickly. “The technology is advancing very quickly. Every day there are announcements about new products and capabilities. For example, agentic AI is an enhanced version of robotic process automation and people will be able to understand it better and build it into workflows.”
Successful AI adoption should start slowly, says PwC’s Lee: “Businesses should walk before they run with AI and identify the use cases that are easy wins. Look at your workflows and see where AI can enhance or streamline a process by either focusing on increased accuracy, efficiency or reduced time or cost. Once you have tackled that, you can move on to more complex processes. The key thing is that the AI adoption journey involves all of your people to help identify where AI can make a difference.”
O’Donovan also emphasises the people aspect. “AI adoption and a supporting data governance culture are more than an IT implementation process; this is about change management and people,” he says. “Organisations, management and employees must understand the AI value proposition and any risks and learn new ways of working and skills to empower them to deliver that proposition.”