AI Strategy Without Execution: Why Progress Remains Out of Reach

EY Ireland research shows that 56% of Irish tech leaders still lack a formal AI strategy and only a small number have clear usage policies in place. The gap between what the prevailing marketing technology narrative would suggest, and execution is startling. Gen AI has generated huge buzz, but if your only answer is GenAI, every problem looks like a prompt.

This isn’t just an Irish story. Recent MIT research found that human-AI collaboration often fails to deliver measurable gains. Even when the AI is highly confident, people tend to override it or use it sub-optimally.  The tech isn’t to blame; the strategy is the problem. If AI tools aren’t embedded in workflows with clear ownership and outcomes, they won’t perform, no matter how advanced they are. [ide.mit.edu]

Section 1: The Weight of the Old

Many organisations are still relying on legacy systems that were built in the past. They would have required significant investment and are complex to change. These platforms, although unwieldy, are systems of record and hold critical data for the organisations they’ve been deployed to.

Now, some of the biggest providers are layering in AI capabilities that boost both usefulness and flexibility. This might mean greater investment to access the new features at a time budgets are tightening.. According to EY’s 2025 Tech Leaders Outlook, only 34% of Irish tech leaders expect their IT budgets to grow over the next two years, down from 46% in 2023. That said, the implementation cost of AI has been lowered through the democratisation of frontier models as components via cloud hyper scalers.  Simply put, AI is cheaper now because big cloud providers offer powerful tools, ready to go.

Section 2: Projects Without a Plot

Across the board (rooms) of Ireland it’s a time of transformation. Teams are rolling out cloud platforms. Automation pilots are running in parallel. Data is being accumulated with urgency. Everyone’s working and busy. But is all this work connected and cohesive? 

In short, no. It’s not.

EY’s Tech Leaders Outlook reveals that 31% of Irish organisations are in the middle of transformation programmes. That sounds promising. Then you look closer. Just 10 percent say AI is central to their strategy. Last year, that figure was 2 percent. The progress is real. But, the hesitation is still obvious.

Most tech leaders in Ireland say they’re not investing in AI. No clear strategy. No long-term plan. They’re interested and they’re watching. But they’re not betting on it. The rationale behind this on-the-fence position is another day’s work but not having an AI Strategy is a problem.

And part of that problem is a landscape full of projects that don’t connect. GenAI is getting plenty of hype as the next big thing. In reality, it’s a broad capability that cuts across functions, not a specialized tool for one sector.  LLM’s are being deployed, but measuring the impact is challenging. Talent is being stretched and budgets are being spent across too many disconnected initiatives.

AI hype has exploded since the launch of ChatGPT and GenAI. But AI is broader than GenAI, and Digital Transformation is broader than AI. AI is a new capability, that should be infused in any modern Digital Transformation Initiative, but to maximise business value requires ensuring that the right technology is used in the right place. 

Section 3: Value Left on the Table

Irish CEOs want to see value before committing to spend. Business value for Digital Transformation initiatives are always challenging, and that challenge has grown with AI. 

A new value calculation model is needed, one that considers AI for the use cases that deliver the best returns. These value calculations also need to consider AI risk. , Under the EU AI Act certain AI Use Cases may require more oversight and carry extra costs. That can reduce the benefits, or return on investment. To get the most from AI, organisations need to initially focus on the use cases that offer strong returns without unexpected regulatory burdens.

The most common AI risks are non-compliance with AI regulations negative impacts to sustainability goals and biased outputs. On an encouraging note, EY’s global risk study found that organisations with real-time monitoring were 65% more likely to achieve cost savings. AI can deliver value, but only when it’s embedded properly within the wider Digital Transformation landscape and business. 

Section 4: The Work Behind the Curtain

An AI strategy typically starts with broad sweeping goals, but it’s important to get past that to the groundwork. That means building out the backlog of use cases, classified by value, organised by a roadmap with an overall mission in mind that reflect the true ambition for the organisation. 

Section 5: Get Started and Make It Real

Execution means aligning IT transformation with business priorities. It means redesigning finance and customer functions so they can actually use AI. It means making sure delivery teams are working from the same assumptions, using the same data, and building toward the same outcomes..

This is where the real work starts.

By addressing these key elements, you can transform even the most elegantly crafted presentation into a powerful roadmap for actionable implementation, ensuring that your AI strategy is both impactful and effective. 

Closing: From Slide to System

AI strategy begins with a clear message. To make it real, teams need to turn that message into concrete use cases measured in business value. Start by identifying processes that slow people down, that can be optimised with AI. Measure the benefit and value and give teams the tools to test, learn, and adjust. 

Make sure the data they rely on is easy to reach and simple to use. Keep the focus on value driven use cases that build momentum. Strategy becomes action when people can point to what’s changed and explain why it matters.

To move beyond PowerPoint, organisations need to operationalise their AI ambitions in three concrete ways:

1. Start Small, Scale Intelligently. Identify specific use cases tied to measurable business outcomes. Prove value early on, then scale with disciplin
2. Invest in Data & People Before Models. Models are only as good as the data they are trained on – and the people who use them. Prioritise literacy, governance, and change management.
3. Institutionalise Learning. AI isn’t a one-time transformation. It is a continuous cycle of learning, experimenting and adapting. Embed feedback loops that keep your strategy alive, not frozen in slides.

These principles form a loop:

Value -> Data -> People -> Learning

They feed each other continuously – that is what turns AI from a presentation into a practice!

If you want to move your AI Strategy from theory to action, connect with EY.