The automation is something you can do relatively easily with the IPFS libraries, API, or CLI.
Ideally your approach to pulling the data from its source into IPFS is somewhat deterministic, so that you get de-duplication and generally more efficient updates.