"People who like X also like..." -- not sure how you would structure this, maybe mostly driven by questions from the audience, but helping people discover games that are similar in some important way to ones they already like, but aren't just the next game from that publisher or whatever, sounds like it could lead to really interesting discussions. Half the work at least would be analyzing what those important factors are -- what can you infer about somebody who like Pandemic, Splendor, and Agricola? I'm imagining a scenario where the person posing the question has to name three games from dissimilar sets as a starting point and the panelists take it from there.

(I don't know if I've ever been to a gaming panel, so apply correction factors as needed.)