Ireland’s next steps for effective AI delivery

Ireland has a strong digital infrastructure public services record, with momentum behind cloud adoption and artificial intelligence. The upcoming update to the National Digital and AI Strategy in 2025 reflects a continued commitment to innovation, not just for economic competitiveness, but for better, more responsive public services. 

To sustain this progress and ensure impact across the economy and society, Irish organisations – particularly in the public sector – will need to take a pragmatic, outcome-led approach to AI adoption. One that prioritises strategic digital skills, cross-sector collaboration, and AI solutions that are scalable and energy-efficient.

Recent Red Hat research shows that nearly all IT managers in Ireland plan to increase investment in both cloud (93%) and AI (95%) this year. The intent is clear. But the challenge is now delivery: how to turn national ambition into measurable outcome, so that all areas of Irish industry can develop an AI toolbox fit for purpose. This is pressing in sectors like healthcare, agriculture and planning, where digital transformation is essential to improving public services, managing environmental resources and addressing infrastructure demands in a growing population. Ireland’s AI use is growing - its application across enterprises grew from 8% in 2023 to 14% in 2024.

Laying the foundations for connected AI 

One of the biggest barriers to delivering progress with AI is structural. Ireland’s public sector is under pressure to improve the speed and quality of service delivery, often while maintaining or modernising legacy systems. Teams are expected to adopt new technologies, collaborate more effectively and respond to rapidly changing citizen expectations. Yet many departments remain digitally fragmented, operating with different infrastructure, governance models and delivery frameworks. Ireland’s public sector has over 900 bodies delivering services at national and local levels, and through AI integration, can offer local governments more autonomy. This shift to digital will be instrumental in offering local bodies a digital foundation that can connect systems and scale innovation on a national level.

According to our research, 96% of IT managers in Ireland say siloed teams are a key challenge to effective cloud and AI adoption as they lead to duplication, inefficiencies and slower innovation. This matters all the more because most public sector opportunities for AI don’t fall neatly within one department. They’re cross-cutting by nature, often spanning health, infrastructure, agriculture and environmental policy. For example, consider a government system designed to anticipate floods in Ireland using data from Met Éireann (weather forecasting), the Office of Public Works (flood defence), the Department of Agriculture (damage/loss & relief), the Department of Health (emergency services), and Local Authorities (public safety information). If each department built and trained its own AI tool, this would produce a fragmented response and wasted resources.

Having a unified platform layer can enable different parts of government to build and deploy AI services in a standardised, collaborative way, sharing best practices and avoiding reinventing the wheel. Red Hat AI provides a secured environment for organisations to build and run their choice of AI models across data centres, on public cloud infrastructure or at the edge. By providing flexibility across hybrid and multi-cloud environments, companies can choose the optimal deployment location for AI models and data. This helps to maintain control and sovereignty of data, meet regulatory requirements and ensure operational efficiency. Think of the benefits if data scientists, developers and IT operations can collaborate on the same platform to build and manage AI workloads. This supports a more sustainable model for delivery: instead of one-off AI pilots that fade when funding ends, teams can reuse across departments or use cases. For example, this can be seen in planning submissions in local government. With older systems now outdated, machine learning can be trained on past decisions to develop an understanding of local policy to complete zoning applications. Cross-functional AI also necessitates cultural and operational change. That means investing in shared digital skills, incentivising cross-agency collaboration and putting in place governance models that encourage openness over isolation. 

Choosing the right-sized AI model for the job

With the updated National Digital and AI Strategy due in 2025, many public sector teams in Ireland will be thinking about shifting from high-level planning to practical delivery.  But not every workload needs – or can support – the largest, most complex AI models. Many of the widely-known AI systems today are built for general-purpose use. They typically demand significant compute resources, rely on cloud-based training pipelines and introduce layers of oversight that may not always suit the needs of regulated or resource-conscious environments. This can make them difficult to apply in practice, particularly in regulated environments with constrained resources.

That’s why many Irish organisations are looking beyond generalised large language models (LLMs) - 84% of IT leaders in Ireland are expanding their approach to include domain-specific solutions, not just one-size-fits-all, reflecting the need for more focused, efficient AI tools. Smaller language models (SLMs) offer a practical route forward, providing similar capabilities to LLMs with reduced computational requirements, designed to carry out specific tasks. Organisations can monitor how the models behave, keep sensitive data within their systems and ensure outputs meet legal and policy requirements. Smaller models also align with Ireland’s digital sustainability goals. They consume less energy, reduce emissions and their size can be adjusted to suit specific local needs. For departments working toward carbon targets, this helps build greener digital services. 

Expanding access to AI 

It is clear that open source is fueling a lot of the rapid innovation in AI. Alongside speed of development, a major reason why open source will continue to be so important for AI is in helping enhance transparency and control for organisations. This applies to platforms and tools as well as the models themselves. Using open source for AI helps make sources and data weights explainable and verifiable  - key factors for building trust and accountability. Open source models continue to be made available, often as SLMs fine-tuned for real-world industry-specific use cases. As mentioned, these smaller models can deliver comparable accuracy to LLMs with lower costs, enhanced data privacy and simplified deployment. 

In Ireland, this approach is already proving valuable. The Department of Agriculture, Food and the Marine (DAFM), working with Red Hat and Version 1, created a machine learning platform called SmartText to scan and classify documents based on sensitivity. By using open tools and reusable components, the team improved system security, streamlined delivery and reduced manual workload, all while retaining oversight of how the model behaves. This model of delivery can benefit smaller departments and local authorities too. By accessing a shared platform, they can apply AI in a way that fits their needs, without having to stand up and manage complex infrastructure themselves. That’s how Ireland can move from fragmented pilots to a more cohesive and practical AI landscape.

Next steps

Ireland’s opportunity is to turn strong intent into practical action by creating a unified AI strategy that is accessible across industry and the public sector. That means building on what already works, sharing tools where possible, and focusing on delivery that’s replicable and cost-effective. Success will depend on whether teams across government and industry can access the models, skills and infrastructure they need, and use them in a way that’s easy to manage and scale. With the right focus on reuse, shared platforms and open technology, Ireland can make AI work where it matters most.