The AI Act is in force and surrounded by myths. The key news for a typical SMB: if you use ChatGPT to write proposals or a chatbot to answer customers, you are not in the high-risk group and no audits await you.
The rules classify AI systems by risk. Banned practices include manipulative social scoring. High risk covers systems deciding about people: recruitment, credit, access to public services. Typical office uses — content generation, document automation, knowledge assistants — fall under minimal risk.
What must you actually do? First: transparency. If a customer talks to a chatbot, they must know it’s AI. Second: human oversight where AI affects decisions about people. Third: basic documentation — which tools you use and for what.
Watch one area: HR. If you use AI for candidate screening, you enter the high-risk category and should review the process with a lawyer. This is the most common SMB touchpoint with the stricter requirements.
Our recommendation: instead of panicking, list your AI tools, their uses and oversight rules in one document. Such an “AI map” takes an hour and covers 80% of what the rules expect from a company your size — while tidying up your rollout too.