What is your goal?Opening a talk session between a Make AI Agent (new module) and an user into a scenario through Telegram.
Session and scenario finish only if the user says « it’s ok 
» or « it’s all good ».
What is the problem & what have you tried?I’m thinking on how to do so :
Is it all I need? Does it work like this?
Essentially yeah - watch messages → AI agent → send message.
What you need to add is some system to track the current conversation. You can store previous messages inside a Data store for reference for example and pull them as well and feed them to the agent so it keeps context. And when the person responds with one of the key words to end the conversation, delete the history as well.
@Stoyan_Vatov thank you.
Do you think that “Conversation ID” field in Make AI Agent module or adding “List Updates” Telegram module as a Tool for AI Agent might be necessary to “keep context” as you said ?
I don’t think either will work well enough. I just store the conversation so far in a data store and then delete it when it’s over.
Hello @Stoyan_Vatov, I’ve been making some tests today, not that good.
Can you please show me how you proceed to store conversations?
I don’t figure out how to concatenate conversations in the same thread (from a talk session opening to the user says « it’s all good »). Watch Updates module doesn’t give opportunity to make precise query.
Have an array of messages and the user identifier in the data store. When a new message arrives, pull the rest of the messages first and feed the whole conversation so far to the AI for reference. Then log the newessayebalong with the new response back to the array.
