I have created a bot that works with the user to create an idea.
I would like the bot to learn from user feedback on the type of ideas that are good and accepted. I would like the bot to learn from every interaction that users have with it and continue to refine its idea creation process and get better over time.
Is this possible? Has anyone done something like this before? How would one go about doing it?
Currently we support end-users up-voting and down-voting responses from the Pickaxes. However, there is not an automatic way that integrates this feedback into your prompts.
However, within a conversation you can a chatbot ask the user question, gather information over the course of a conversation, and then use that information to generate a result. I suggest checking out the “course creator” chatbot builder.
Hi Mike, what happens to those up votes? Is there a possibility to get them in the data we download, it could help us build fine tune model 
I think you could pass the user input via AI to a server by instructing the AI to ping https//your-server/receiver.php?inputWas=urlencode(user input) which you could then capture and store for usage as a document within the same AI documents?
In other words, the user asks for an idea about cooking brown rice and the AI makes a suggestion (ping the question, ping the proposed suggestion) and have the AI ask the user if the response is a good idea…(Ping their answer, use another AI to classify it, store to a document, make sure that document is part of your AI training documents)
Why did the upvotes disappeared?