ML algorithm analyzes the words and phrases in the context of the entire message. This allows for a more intelligent determination of whether a flagged word may be considered profane by the requester.
Ability to configure a profane message violation threshold for agents based on the number of profane messages sent. Agents will receive updated warnings as they approach the violation threshold. If they meet or exceed the violation threshold, the chat supervisor will be alerted with the option to join the conversation.
Customers can switch between ML-based and keyword-based detection. Customers have the ability to curate their own custom keyword list.
Chat supervisors can view a log of all messages with flagged words and phrases and the names of the agents who sent them.
All messages with flagged words and phrases are not displayed to the requester but are highlighted on internal transcripts.