The AI Act gives HR more time. Trust cannot wait.

A candidate disappears from a shortlist. A recruiter sees a match score of 42%. A manager accepts the ranking and moves on.

Nobody deliberately decided to let AI make the hiring decision. Yet somewhere between the software's recommendation and the next interview invitation, that is effectively what happened.

This is where the AI Act becomes a leadership issue for HR.

As Europe moves into a new phase of AI Act enforcement, one distinction matters: not every obligation follows the same timetable. The European Commission's updated schedule places the application of the high-risk rules covering areas such as employment at 2 December 2027. An August 2026 enforcement milestone should not be confused with all recruitment-related requirements becoming applicable at once. (European Commission: regulatory framework on AI)

That gives organisations more preparation time. The decisions their systems influence, however, are happening today.

AI can help HR reduce repetitive work, organise information and give recruiters more time for meaningful conversations. But when a tool influences who gets seen, interviewed or offered an opportunity, efficiency is only part of the picture.

People also need to know who is responsible.

The change can arrive through a software update

AI adoption does not always begin with a strategy meeting. It can arrive as a new feature in an applicant tracking system, a candidate recommendation or an automatically generated assessment.

Each feature may look like a small improvement. Together, they can change how an organisation makes decisions about people.

The AI Act identifies recruitment uses such as analysing and filtering applications and evaluating candidates among its high-risk use cases. The intended purpose and specific use matter; not every AI feature used by an HR team belongs in the same category.

For HR leaders, the practical starting point is understanding what happens in their own processes.

Three questions make that conversation concrete.

1. Where does AI influence someone's opportunity?

Follow the candidate journey from the first search to the final decision. Identify where software filters, scores, ranks, summarises or recommends.

Then ask what happens next.

Does a recruiter examine the full application, or only the summary? Can someone with a low score still reach an interview? Are candidates excluded before a person sees their profile?

A useful inventory connects each tool to the decision it influences and the person responsible for that decision.

2. Can the person reviewing the output actually challenge it?

"Human review" sounds reassuring. Its value depends on what the reviewer can do.

Consider a recruiter processing hundreds of applications against a tight deadline. A score appears on screen, but its basis is unclear. Overriding it requires extra work.

The person is present. Meaningful oversight may still be missing.

Make review workable. Give people enough context to question a recommendation, clear authority to override it and a route to raise concerns. Managers should make it clear that challenging an AI output is part of doing the job well.

3. What happens when the system gets it wrong?

Imagine a strong candidate being ranked poorly because their experience follows an unusual career path.

Who notices? Who investigates? Can the decision be reconsidered? Who contacts the supplier if the same problem keeps appearing?

Agree on that process before the first complaint. A named owner and a clear escalation route make accountability practical.

Turn preparation time into progress

Bring HR, IT, legal and the relevant process owners together. Start with one workflow, such as recruitment, and map where AI influences it.

Ask suppliers for specific information: what the feature does, its intended use, known limitations, available oversight controls and their approach to the applicable AI Act requirements. A broad assurance that a product is "compliant" cannot answer every question about how your organisation uses it.

Then explain the agreed approach to the people working with the tools. Show them where AI helps, where their judgement matters and how to raise a concern.

At Happy2Change, we see this as a strategic HR responsibility: connecting technology choices with leadership, culture and the everyday experience of work.

The opportunity is to use the time available to build habits people can trust.

Start with one question at your next leadership meeting:

"If AI influenced a decision about someone here today, could we explain what happened—and who took responsibility?"

If the answer is unclear, you have a useful place to begin.

Talk to Happy2Change about making those responsibilities clear and helping your people work confidently with AI.