In the print dialog choose “Save as PDF”.← Back to the article
ISSUE OCTOBER 16, 2026BONITA SPRINGS · FLORIDA · WORLDWIDE
THE KESS TIMESBUSINESS BUILDING × TRANSFORMATION × AI × MARKETING
#thatskess

SECTION: AI × OCTOBER 16, 2026 × 4 MIN READ

EU AI Act 2027: What You Really Have to Do as an Entrepreneur. And What You Don’t.

BY THE KESS FACTORY

Hardly any law has caused as much uncertainty as the EU AI Act.

And hardly any has been explained wrongly so often.

Some say: “Doesn’t affect me, I only use ChatGPT.”

Others say: “Better no AI at all than get into trouble.”

Both are pretty un-KESS.

Time for some straight talk. What already applies? What’s coming in 2027? And what do you as an entrepreneur need to do now, specifically?

A quick note up front: this article is an overview and not legal advice. For your specific case, please talk to your lawyer.

WHERE THINGS STAND IN ONE SENTENCE.

With the so-called Digital Omnibus, the EU has pushed back the strictest obligations.

But not scrapped them.

If you’re now thinking “Great, then I’ve got ages”, read on for a moment.

POSTPONED ISN’T FORGOTTEN.

PICTURE OF THE ISSUEFresh from the KESS FACTORYThat's how Bradley sees it.

WHAT ALREADY APPLIES. YES, TO YOU TOO.

SINCE FEBRUARY 2025: AI LITERACY AND CLEAR BANS.

Certain AI practices are off limits. Manipulative systems or social scoring, for example.

And then there’s one obligation that affects almost every business: AI literacy.

If you use AI in your business, you must make sure the people working with it know enough about it. What can the tool do? Where are its limits? Which data has no business being in it?

Sounds like bureaucracy. But it’s really just common sense.

SINCE AUGUST 2025: RULES FOR THE BIG MODELS.

The providers of large AI models have their own obligations. That mainly affects the tech giants. You as a user only indirectly.

NOW ALSO: TRANSPARENCY.

The transparency obligations have applied since August 2026. Put simply: people must be able to recognize when they’re talking to an AI. And certain AI generated content must be recognizable as such.

So the chatbot on your website is welcome to be charming. It just mustn’t pretend to be Sabine from sales.

AI CAN HELP. BUT NOT IN DISGUISE.

WHAT’S COMING IN 2027: HIGH-RISK AI.

After the postponement, the obligations for standalone high-risk AI systems apply from December 2, 2027. AI in regulated products, such as machinery or medical devices, follows in August 2028.

And what does “high-risk” mean?

For example, AI in recruiting. In evaluating employees. In lending decisions. Or in critical infrastructure.

Anyone running something like this needs risk management, documentation, logging and human oversight, among other things.

That’s not a weekend project.

IF YOU WAIT UNTIL FALL 2027, YOU’RE STARTING TOO LATE.

DOES THIS EVEN AFFECT ME?

The honest answer for most small and medium-sized businesses: partly.

If you use ChatGPT for copy or have a chatbot on your website, you’re not a high-risk case. Breathe.

But AI literacy and transparency affect almost everyone.

And if you use AI in recruiting, have it pre-sort applicants or evaluate employees automatically, you should take a very close look now.

NOT EVERY AI IS HIGH-RISK. BUT EVERY AI NEEDS A PLAN.

YOUR KESS TO-DO LIST FOR 2027.

1. AI INVENTORY. WHAT’S ACTUALLY ALREADY RUNNING?

Which AI tools do you use? For what? With which data? And who actually decided that at some point?

Spoiler: in many businesses, employees have long been using AI the boss knows nothing about.

2. CLASSIFY THE RISK.

Does any of it fall into a high-risk area? If yes: turn it into a project. If no: stay relaxed, but document it.

3. TRAIN. AND PROVE IT.

A short internal training plus a clear AI policy. What’s allowed? What isn’t? Which data stays out? Write it down. Done.

4. CHECK TRANSPARENCY.

Chatbot labeled? AI generated content marked where necessary? Quick check, big effect.

5. PLAN HIGH-RISK AS A PROJECT.

If you’re affected: start now. Not in November 2027.

ORDER ISN’T A BRAKE. IT’S THE TURBO.

AND THE PENALTIES?

They’re hefty. For prohibited practices, up to 35 million euros or 7 percent of worldwide annual revenue. For violations of high-risk obligations, up to 15 million euros or 3 percent.

There are reliefs for small and medium-sized businesses. But that’s no free pass.

CONCLUSION: NO REASON FOR AI FEAR. A REASON FOR AI CLARITY.

The AI Act is no reason to do without AI.

It’s a reason to use AI deliberately.

If you know which tools you use, train your people and document a few things properly, you’re already further than most.

And you can use AI for exactly what it’s there for: less routine, more room for what really counts.

SMART IS GOOD. COMPLIANT AND SMART IS KESS.

READ ALSO: AI agents: 7 jobs for your digital employee · Accessibility on your website

#thatskess

Want to set up AI in your business safely and sensibly? In the SCALE IT Way, we build AI systems that work and fit you. FIND YOUR KESS WAY

Read online: https://kessfactory.com/en/kess-times/eu-ai-act-2027-en/WE BUILD BUSINESSES FOR THE NEW TIME. · kessfactory.com