Improving Spark Bot: Addressing Spam and Response Quality
Summary
The CommunityOne Builders Help server is actively addressing feedback regarding the Spark bot. This update focuses on mitigating spam caused by excessive tagging and long-form responses, while improving the consistency of smart responses. Developers are working on new spam control frameworks and formatting solutions to ensure the bot enhances rather than disrupts community conversations.
First issue is that Spark can be tagged on general chat excessively and hence disrupt a regular community flow.
Second, the smart response is not very good. It doesnt answer member questions consistently, and sometimes it answers in a way that is disrupting another set of conversations. Maybe a better solution is to use a bot embed message to clearly differentiate in the smart response mode?
How about ignore channel/role/user function
so this is being actively worked on in this part of spark and there should be som resolution
but a basic framwork is that we allow you to set up spam control, if a user chats x message too much consetiviely in channesl, then we direct traffic to a chnanel that u specify
X message doesn't matter when a single message can contains whole huge chunk of text occupying full screen
Like its still spam
Imagine
I meant spark's response can be sometimes very long which i consider as spam
