EU AI Act Delay: More Time Isn't a Strategy

European lawmakers reached a provisional deal to delay the EU AI Act's high-risk rules. Obligations for standalone high-risk systems move from August 2026 to December 2027, and rules for AI embedded in regulated products move to August 2028. The deal also adds new prohibited practices.

Critics call it caving to industry. Supporters call it realism, since the technical standards companies need to comply weren't ready. Either way, the practical effect is the same: you have more time. The question is what you do with it.

What the EU AI Act delay changes

The schedule, not the destination. The high-risk categories still include AI used in employment decisions, access to essential services, and other areas that touch customers and workers directly. The obligations around risk management, documentation, data quality, human oversight, and transparency are still coming.

The rest of the Act keeps moving. Prohibited practices and the rules for general-purpose AI models weren't paused, and the deal adds to the list of banned uses.

A delayed deadline is still a deadline. It just rewards the companies that use the time.

Why this matters beyond Europe

If you serve European customers or employ people there, the Act applies to you wherever you're based. And even if you don't, the EU's framework is becoming a reference point for how regulators elsewhere think about AI risk.

Contact centers sit closer to the high-risk line than many leaders assume. AI that scores, routes, or prioritizes people, or that supports decisions about access to services or employment, deserves a careful look.

How to use the extra time well

Map your AI uses against the high-risk categories now. Know which systems might qualify and which clearly don't, and document why.

Build the boring foundations. Logging, documentation, human oversight procedures, and data governance take longer to stand up than any compliance team expects. Start them while there's no fire.

Watch the standards. Much of the delay exists because the technical standards weren't finished. When they land, they'll define what "compliant" means in practice. Have someone tracking them.

Companies that treat this as a pause will spend late 2027 scrambling. Companies that treat it as a runway will be ready, and will probably have better AI operations along the way.

Regulators gave you time. Don't give it back by waiting.

Source: Hogan Lovells, EU legislators agree to delay for high-risk AI rules