The EU AI Act deadlines may have shifted, but the compliance countdown has not stopped.
The rules for standalone high-risk AI systems, including those used in recruitment, education, healthcare and law enforcement, have been pushed back from August 2026 to December 2027.
High-risk AI systems built into regulated products, such as medical devices and certain types of machinery, now have an extended transition period until August 2028.
But that does not mean organisations can ease off.
Many other AI Act obligations are still moving ahead, including transparency requirements and rules around AI-generated content. The AI literacy obligation has already been applicable since February 2025, requiring organisations to ensure that people operating AI systems have the appropriate skills and understanding.
Here in Ireland, the government has designated 15 national competent authorities under the AI Act and is establishing a National AI Office to coordinate implementation.
So, while some of the toughest deadlines have moved, there is still real work to do.
Here are seven practical steps your organisation should be taking now.
1. Create an AI Inventory
You cannot govern what you do not know exists.
Start by asking every department a simple question: What AI tools are you using?
Marketing teams may be using ChatGPT. HR could be trialling AI recruitment tools. Finance may have AI features built into accounting software, while product and UX teams could be using AI-powered analytics and research tools.
Much of this may already be happening without formal approval. This is often referred to as shadow AI.
Before you can think seriously about compliance, you need visibility across your organisation.
2. Classify Your AI Use Cases
The EU AI Act takes a risk-based approach.
Many everyday productivity tools will fall into the low or limited-risk categories. These could include writing assistants, summarisation tools and internal AI copilots that are not making high-stakes decisions.
However, some AI uses in areas such as recruitment, education, healthcare, law enforcement and public administration may be classified as high-risk.
These systems can influence who gets a job, who accesses essential services or how important public decisions are made.
Even though some high-risk obligations have been delayed, organisations still need to understand where their AI use cases sit on the risk spectrum.
3. Establish Clear AI Governance
Who is responsible for AI in your organisation?
Who approves AI tools? Who manages your AI policies? Who monitors suppliers and keeps track of regulatory changes?
Without clear ownership, AI adoption can quickly become fragmented and difficult to manage.
Whether you establish an AI steering group, AI governance board or digital risk committee, the goal is the same: put names against responsibilities.
Your organisation should know exactly who is accountable for AI-related decisions.
4. Update Your Policies
Most organisations already have policies covering social media, cybersecurity and data protection.
But many still do not have an AI policy.
Your AI policy should clearly explain:
- Which AI tools employees are allowed to use
- How organisational and customer data should be protected
- When human oversight is required
- How AI-generated content should be reviewed
- Which AI uses are prohibited
For example, employees should know whether they are allowed to enter confidential client information into publicly available AI tools.
A good AI policy should not make it harder for employees to use AI. It should give them the confidence to understand what is safe, responsible and appropriate.
5. Train Your People
This could be one of the most important investments your organisation makes.
AI tools change constantly. Strong AI skills and understanding last much longer.
Employees need to understand issues including bias, hallucinations, copyright, confidentiality, security and human accountability.
They also need to know when to escalate a decision and when human judgement should override an AI-generated recommendation.
AI literacy is not simply a nice-to-have. It is already an obligation under the EU AI Act.
Organisations should treat AI training as both a compliance requirement and a core business capability.
6. Review Your AI Suppliers
Your AI compliance responsibilities do not stop with the tools your organisation develops or directly manages.
AI is increasingly embedded in third-party software, often without users even thinking of those systems as AI tools.
Organisations should be asking suppliers important questions:
- How does the AI system work?
- What data does it process?
- Where is that data stored?
- How is the system governed?
- What documentation can the supplier provide?
If a supplier cannot provide clear answers, that should be treated as a governance concern.
Your organisation still needs to understand how third-party AI systems affect your employees, customers and stakeholders.
7. Document Everything
Accountability is a central part of the EU AI Act.
If a regulator asks how an AI-related decision was made, your organisation will need evidence, not just assurances.
That means maintaining records of:
- AI training
- Policies and procedures
- Risk assessments
- Governance meetings
- Supplier reviews
- AI use case classifications
- Important decisions and approvals
Good documentation allows your organisation to demonstrate that AI risks are being actively and responsibly managed.
Why the EU AI Act Delay Is an Opportunity
It would be easy to see the delayed deadlines as a reason to ease off on AI compliance.
Organisations should do the opposite.
Those that use this additional time to create an AI inventory, train employees, establish governance and improve their policies will be better prepared when the remaining deadlines arrive.
They may also gain a competitive advantage.
AI governance is increasingly becoming part of the same conversations as cybersecurity and GDPR. Customers, governments, boards and other stakeholders want to know that organisations are using AI responsibly.
For communications professionals in particular, this presents an opportunity to lead.
Trust, transparency and stakeholder engagement are already central to our work, and they are exactly what responsible AI adoption requires.
The deadlines may have moved, but the direction of travel has not.
The question is: is your organisation using this extra time to get ready?
If you want to build your AI literacy and skills, I can help
 Grab AI Playbook on Amazon, the perfect podcast companion.
 Subscribe to Human in the Loop, my LinkedIn Newsletter
 Take the CPD Accredited AI Playbook Diploma
  Book a call at digitaltraining.ie
Sources
European Commission, Regulatory framework on AI (application timeline and Digital Omnibus changes, updated 11 May 2026): https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-aiÂ
Department of Enterprise, Tourism and Employment (Ireland), Ireland leads the way in EU AI regulation, 16 September 2025 (15 national competent authorities and National AI Office): https://enterprise.gov.ie/en/news-and-events/department-news/2025/september/20250916.html
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
Connect with Joanne on LinkedIn.
Â
