EU AI Act: What Irish businesses need to know amid new rules

The EU AI Act came into force on the 2nd August 2026. AI has moved quickly from experimentation into day-to-day business activity. In many industries, it now provides the data behind key decisions, supports how customers are managed and conducts preparatory work ready for review by your team. However, the adoption level across businesses and industries varies significantly.

In most cases, AI hasn’t been introduced as part of a single or official rollout. It has built up over time; a recruitment tool here, a marketing platform there, often introduced to solve specific problems rather than as part of a wider plan. As a result, it’s not always clear where AI is being used or how much influence it has across the business.

That all changed for EU businesses in August. The changes bring new expectations around how AI is understood, managed and, where needed, explained.

This article looks at what the EU AI Act covers, what has changed from August, and where businesses are most likely to feel the impact.

What Irish businesses need to know about the EU AI Act

What is the EU AI Act?

The EU AI Act is the first comprehensive attempt to regulate artificial intelligence across the European Union. It takes a risk-based approach, meaning the level of regulation depends on how AI is used and the level of risk involved.

At a high level, AI systems are grouped into four categories:

• Unacceptable risk – systems that are prohibited altogether (for example, AI used for social scoring or certain forms of biometric surveillance)
• High risk – systems that can be used, but are subject to strict requirements around data, documentation, human oversight and monitoring
• Limited risk – systems that require transparency, such as making it clear when users are interacting with AI
• Minimal risk – systems that can continue to be used with minimal regulatory impact

Most businesses will sit somewhere in the middle, particularly where AI supports internal processes or customer interactions, such as chatbots.

How the technology is used will determine how it is treated under the Act. The same tool can fall into different categories depending on its role. For example, using AI to draft marketing content is likely to be low risk, while using similar technology to screen job applicants or assess customer eligibility would be treated very differently.

The focus is on situations where AI is influencing decisions that affect people, access to services or outcomes such as employment or credit. In those areas, there is an expectation that the system can be understood, explained and properly overseen.

What has changed since the EU AI Act came into force?

The EU AI Act is being introduced in stages, but 2nd August 2026 is when it began taking effect.

From this point, businesses are expected to be able to identify where AI is being used and explain how it is being applied. This includes:

• Knowing which tools or systems involve AI-powered processes.
• Understanding how those systems would be classified under the Act.
• Having appropriate oversight in place, particularly where AI supports decision-making.

Not all requirements apply immediately, but the expectation that AI use can be identified and explained is already there.

That’s a shift in itself. In many businesses, AI has been introduced gradually, sometimes at team level, without being tracked or reviewed centrally.

The difficulty comes when there is a need to explain how decisions are being supported, particularly in areas like recruitment, marketing and customer operations. You should, as a business owner, know where AI is being used, how it is being used and how your team is conducting appropriate human oversight.

AI and your finance function

While AI is used across many areas of a business, one of the most common and practical applications is within finance and accounting systems. Tools like Xero, Dext and similar platforms already use AI to automate processes, categorise transactions and flag anomalies.

In many cases, these tools are influencing outputs that feed directly into financial reporting, VAT returns and management decisions.

That’s where the focus starts to shift.

Where AI is shaping financial data or outputs, there is an expectation that those results can be explained clearly and, if needed, justified.

For example:

• How transactions are categorised.
• How expenses are matched or reconciled.
• How VAT is calculated or flagged for review.

These are areas where AI can add real value, but also where errors can have a direct financial impact if they go unchecked.

This is often where things become less straightforward. The process works, the outputs look right, but the underlying logic isn’t always clear or reviewed.

In these situations, it’s not enough to rely on the system. There needs to be a clear understanding of:

• What data is being used.
• How outputs are being generated.
• Where human review fits into the process.

Used well, these tools improve efficiency and accuracy. Used without oversight, they can produce incorrect results quickly and at scale.

We’ve written previously about how AI is changing day-to-day accounting,  worth a read if you want the practical upside alongside the compliance picture.

What should businesses be doing now to prepare?

The focus now should be on understanding how AI is being used within your finance function, particularly where it feeds into reporting, compliance or decision-making.

From there, a few areas usually need attention:

• Look at how cloud accounting tools like Dext, Xero or similar platforms are being used- These systems often automate coding, matching and data entry. It’s worth understanding where those outputs are being relied on without review.
• Review how financial outputs are generated- If AI is supporting reconciliations, VAT returns or reporting, there should be clarity on what’s driving those outputs and how reliable they are.
• Make sure there is appropriate review in place- Automation can reduce manual work, but it shouldn’t remove oversight. There should be clear points where outputs are checked before being finalised or submitted.
• Be clear on who understands the system- In many businesses, one person manages the setup and understands how the tool works. That can create risk if that knowledge isn’t shared or documented.
• Be confident explaining your process- If asked how figures have been generated or decisions supported, there should be a clear and consistent answer, rather than having to work backwards to piece it together.

The goal is to have a clear, central view of how AI is being used within your finance processes, where it is influencing outputs, and how those outputs can be explained.

How can AAB help?

At AAB, our digital advisory team helps clients get both sides of this right – using AI to improve accounting and reporting, while making sure the oversight, documentation and controls the EU AI Act expects are actually in place. We’ve covered the practical side of this before, we’re firm believers that AI in accounting should be empowering your teams, not replacing them.

One area where AI is already making a difference is in accounting, when it’s used with the right oversight and control.

At AAB, our digital advisory team uses AI to support the advice we give to clients. Client data remains secure, outputs are reviewed, and technology is used to inform decisions, not make them.

Used well, AI improves accuracy and frees up time to focus on what matters most. Used poorly, it just gets you to the wrong answer faster.

With the EU AI Act now coming into force, understanding how AI is being used within your finance function, and whether the right controls are in place, is becoming increasingly important.

If you’d like to discuss how AI is being used within your business, or where additional oversight may be needed, please contact Graham Dyer or your usual AAB contact.