A customer who sends an angry or threatening reply to a text conversation deserves a different response to one who is simply asking when their order will arrive, and getting that distinction right protects both a support team and the people they are messaging. Missing the difference, or only catching it after someone has already read the message, is where the damage from an abusive reply tends to happen.

Tisane's job is reading language rather than sending messages. It is built to pick up cyberbullying, hate speech, sexual advances, obfuscated profanities and other criminal or abusive content hidden in a piece of text, which matters the moment that text becomes a reply sitting in an SMS inbox. It works on the words themselves, not on who sent them, how many messages have piled up, or which staff member is meant to be watching the queue that day.

The practical effect is that a nasty reply gets treated differently long before a staff member scrolls past it in a busy inbox, so the people replying to a text are met with judgement calls that have already had the worst of the language filtered out first. Connected to a wider workflow through Zapier, that kind of language check can sit alongside the ordinary business of sending and receiving texts rather than replacing it.