You can always lookup all the received TLVs and figure out what's data and what's not. But at the end of the day applications that use this solution for transmitting data over LN should know what keys they transmit signatures/data over. Just like existing web services use specific ports.

A reason you want each application to listen to specific TLV keys is some kind of filtering. You could have multiple things running on top of your node, and one way to indicate the intention of a specific TLV is by its key (some form of "type").

For example if we have a messaging application then it will know that TLVs intended for messaging are placed in the key X. Other applications could be running in parallel and use other TLV keys for their traffic.