Right now, when you get a call transcript, you can see all of the spoken words, but you can't see who was speaking. Ideally, you'd be able to see the texts spoken by each person in the call.

George: Hi there

Sam: Hi George, how are you?

George: I'm good...

etc...

In order to make the transcript helpful right now, you have to know who is actually speaking. Otherwise, you have to intuit what was going on in the conversation, and that's not as easy at times, and doesn't provide an LLM the kind of information necessary to pull really good call notes after the fact