My understanding is that it's usual to create a pitch-deck, containing various elements of what you're pitching.
(Rami Ismail recently posted
this template--and it may be worth looking at other game-dev articles on the same site--while publisher Raw Fury some time ago posted
their own template.)
Regarding a demo specifically, some publishers actually request one as I recall. And while I daresay that a video may not be a bad idea, having a prototype-demo seems likely to be a good one: those that want to try the game can, and those that want to move on needn't play it.
And indeed, looking at
this article on the former of the two sites linked above, I see this advice:
Personally, I believe pitching is something that you can (& should!) start doing after the prototype, as long as you have a number of mockups available to communicate the atmosphere of your work.
The reason is simple: if the Prototypes are what define whether the game should be made, it's a very good way to see if anyone is interested in the game in the first place.
(What that article argues
against, I believe, is going all the way to the "vertical slice" stage of development before pitching, instead pitching with a
prototype.)
Finally, it might be worth your looking at the following article--granted that I don't know how up-to-date it is:
How to guarantee your pitch will be rejected