The system inspects draft text against competitor domains using a regular expression that matches word-like boundaries while permitting leading dots for subdomains.