I want to create a context window for the user to send new messages to my bot (WhatsApp) and have a conversation.
I’m building a message flow where the user sends a single message, and an OpenAI agent classifies it, directing the user to a specific flow. However, now my requirement is that the user needs to interact multiple times within one of the flows to build a document. Example use case:
User: I want to create document X
Bot message: Perfect! Please provide the title.
User: [Random name]
Bot message: Question 1 - “What was the most important thing that happened today?”
User: [Random answer]
The problem lies with these new messages. Currently, each message sent to the bot is treated as a new request, with no memory. How can I implement this in the best way? Is there already a solution for this?
Here is my workflow + I change the values in the nodes (security).
Information on your n8n setup
- n8n version: 1.71.3
- Database: Supabase
- Running n8n via (Docker, npm, n8n cloud, desktop app): Desktop App