A polished scam can still be a scam.
SARS specifically warned in 2026 that criminals were using AI to generate professional-looking email templates.
That makes grammar and spelling weaker indicators than they used to be.
Instead, examine the request.
Does it ask you to click a link? Does it demand credentials? Does it involve money? Is it unexpected? Can you verify it through the official website?
Those questions remain useful even when the message is perfectly written.
The goal is not to become an expert at spotting AI-generated text.
The goal is to stop trusting a message merely because it looks professional.