Report by AI Ireland’s Mark Kelly shows strong momentum toward applied AI, tempered by legacy-system hurdles and cautious data governance.
Irish businesses are rapidly moving artificial intelligence (AI) from experimentation into day‑to‑day operations, according to a new national survey authored by Mark Kelly, founder of AI Ireland.
The report draws on responses from 130 senior leaders and finds that 41% of organisations are already in early production and a further 33% in broad adoption. Only 2% have yet to begin using AI, signalling that the question for most companies is no longer whether to implement the technology but how to scale it safely.
“The 56% of leaders favouring controlled use are essentially already following the right model, however they just need to formalise it”
Commercial priorities dominate the near‑term agenda. Cost optimisation is the most common objective, followed by engineering assistance and predictive maintenance designed to identify failures before they occur. Leaders also highlight specialised use cases around lead generation, process efficiency and contract intelligence, reflecting an appetite for tools that deliver measurable returns.
Fundamentals of AI for business
Responding to questions from ThinkBusiness, Kelly outlined a set of fundamentals for any organisation planning to expand its use of AI.
“Before starting any AI project, I would say, audit your existing IT systems to identify what can and can’t connect with new AI tools. Clean up your data, assign a dedicated executive to own the integration process, and define clearly what business problem you’re actually trying to solve. Companies that skip this preparation phase are the ones that stall mid‑project.”
Integration remains the most significant obstacle. Almost a quarter of respondents cite the difficulty of connecting modern AI systems with legacy platforms as their primary blocker, while shortages of skilled staff and data security concerns also feature prominently.
Leaders note that successful projects are those that build user trust by demonstrating clear business value and that teams often need several iterations before establishing a stable production workflow.
Caution shapes how organisations use generative AI with their own data. A majority prefer controlled deployment, with 56% favouring strict governance, while only 15% say they are production ready and another 15% limit such tools to sandbox experimentation.
Kelly argues this posture should be formalised. “The 56% of leaders favouring controlled use are essentially already following the right model, however they just need to formalise it. Test in a sandbox first, prove the business value, then scale with guardrails in place. Draft simple ethical guidelines early, pursue ISO 42001 certification as your compliance framework, and make governance part of how a project earns its way into production rather than a separate layer bolted on afterwards. That way you stay safe without grinding innovation to a halt.”
Operational experience captured in the survey reflects both optimism and realism. Companies report strong gains from code generation for app modernisation, and widespread use of Microsoft Copilot to prepare meetings and manage supplier contracts.
At the same time, delivery often takes longer than planned and requires stricter code reviews to catch security and quality issues. Leaders also highlight the risk of adding features that confuse users rather than help them, a trend amplified by vendors eager to attach AI labels to their product roadmaps.
Short‑term priorities lean toward early‑warning systems and efficiency. Predictive alerts top the 90‑day roadmap, alongside cost optimisation, incident support and contract intelligence. More than a quarter of organisations are developing bespoke tools tailored to their sectors. The report warns that progress against these immediate goals will depend on addressing integration constraints, which 25% of leaders identify as their most pressing hurdle.
The survey sample skews toward larger companies, with more than two fifths of respondents working in organisations employing between 1,001 and 5,000 people. Senior leaders dominate the responses, giving the findings a corporate‑enterprise perspective on adoption.
Kelly’s report concludes with a pragmatic roadmap. He urges companies to audit legacy systems, map data lineage, build secure sandboxes for generative AI testing and promote pilot projects only after they show clear business value.
Top image: Mark Kelly, AI Ireland
-
Bank of Ireland is welcoming new customers every day – funding investments, working capital and expansions across multiple sectors. To learn more, click here
-
For support in challenging times, click here
-
Listen to the ThinkBusiness Podcast for business insights and inspiration. All episodes are here. You can also listen to the Podcast on:
-
Spotify
-
SoundCloud
-
Apple
