I’m working on a flow to help out accounting department, to classify emails,
“invoice”, “payment reminder”, “account statement” ect.
however, the difficulty is that the emails are quite varied, and I need to classify them based on the entire message and all attachments, a email can have a question in the text and include the original invoice, so that it should NOT be classified as an “invoice”, but as a “question”
I’m having trouble getting an ai “fed” with all the info it needs to classify the:
- email body
- pdf emails
- pdf SCAN emails
- image attachments
and then give a ‘combined’ classification.
my current attempt is to convert all attachments to text and saving them in a qdrant vector store, and then “ask” the AI, but as I’m building it, it feels like the wrong approach.
- n8n version: 1.77.3
- Database (default: SQLite): Postgres
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
- Operating system: MacOS