Most businesses have barely noticed.
Here’s why that changes in August 2026.
– High-risk AI obligations hit on 2 August 2026.
– Fines of up to €35 million or 7% of global turnover activate.
– Finland already has enforcement powers operational.
The businesses that will get caught out aren’t the ones ignoring AI. They’re the ones that deployed it without thinking about what category it falls into.
The truth is:
40% of enterprise AI systems have unclear risk classification.
Neither the businesses running them nor the regulators overseeing them can agree on where they sit.
And the technical standards you need to prove compliance?
They may not be published until Q4 2026.
After the deadline.
There is a plot twist, though.
The European Commission proposed softening the Act in November 2025.
The Digital Omnibus package would defer high-risk obligations to December 2027 and cut admin burden by 25%.
But that proposal is not law yet.
You cannot plan around a maybe.
For UK businesses, there’s additional complexity. The UK has no standalone AI law- yet. But if you serve EU customers or use EU data, the Act applies to you regardless.
Three questions every business should be asking right now:
1️⃣ Do we have an inventory of every AI system we’re running?
2️⃣ Have we classified each one by risk level?
3️⃣ Have we started our conformity assessment? (It takes 6-12 months.)
If the answer to any of those is no, August 2026 will arrive faster than you expect.
Have you started your AI Act readiness assessment?
