There is a category of language that has become so associated with bad sales outreach that prospects now use it as a filter for who to ignore.
"Hope this finds you well." "I would love to connect and explore synergies." "Would you be open to a quick 15-minute call?"
These phrases did not start out bad. They became bad through overuse. And AI made the problem significantly worse by making it trivially easy to generate high-volume, low-quality outreach that all sounds the same.
Suppress, don't filter
Most AI writing tools try to fix bad output by filtering it after generation. Vaunzo takes a different approach: active suppression at generation time.
Vaunzo maintains a registry of phrases that are blocked from appearing in any generated message. Not flagged. Not suggested for revision. Blocked. The generation system is instructed to avoid them entirely, which forces the model to find more specific, more human language.
Paired with a positive constraint
Suppression alone is not enough. Vaunzo also enforces a positive constraint: every generated message must end with a genuine question about the prospect, something specific to their role, their background, or their situation.
Not "Would you be open to a call?" A real question that invites a real response.
Together, the suppression layer and the positive constraint produce outreach that reads like it came from someone who did their homework.
This is what separates AI-assisted outreach that works from AI-generated noise that does not.