AI limit number of messages per visitor

Since you have packages by number of messages it would be grate to avoid abuse AI from clients and messages. Like if someone can’t find answer with 50 messages with comunication with AI then alert agent and transfer to agent.

Hi djordjeb,
Thank you for your suggestion to improve our service. The idea of transitioning difficult AI conversations to human agents after a set number of messages is indeed valuable.

Your proposal aligns with our commitment to customer satisfaction and efficient service. I’ll share this with our development team for further consideration.

Let us know if you have any other suggestions.

I trid something in base prompt but there is no resuls.
Something like:

  • On second message from user write him that he can contact support trough email info@email.tld and after that do not send him any message.

AI still answering and tried any version of that but still without succes.
Anyone tried and succeeded?
So idea is to answer [number] of times and after [number]+1 ignore all future messages from user.

It can actually count number of questions but not good.

Hi there,

It is not possible to make AI ignore user messages as it is set to provide a reply to every message it receives.

You could potentially try something like:

  • At the beginning of conversation, always greet visitor and welcome them to our website.
  • Once user replies to your greeting, reply with ‘‘You can contact support at’’
  • Once you have provided user with email address that they need to contact, do not provide any additional responses. If user sends additional inquiries, tell them that you will not be able to help them and that they need to contact a human member of our team via email without repeating yourself.

Also, please note that while counting messages might work at the start of the conversation, the longer it goes on the more likely it is that AI will lose track of the count due to the way the conversation so far is summarized in the eyes of AI.

That is good but with limitation we can avoid abuse since competition is not fair.

Look at this screenshot example, where even regular user who is bored he can abuse AI.