Doing facial motion capture for every NPC in the game is impossible. You have to remember that we're talking about recording massive amounts of lines with numerous actors across different localizations, cooperating with studios all over the world for sessions, and doing VO pickups all the way until the very last moments sometimes. Using JALI we're able to successfully overcome many difficulties, for example logistical ones, that this undertaking would otherwise pose with regular facial motion capture.
When I think about this relationship between CD PROJEKT RED and JALI in general, it reminds me of the quote from John Lasseter who once said that the art challenges the technology, and the technology inspires the art. In my opinion, what we've managed to achieve with Cyberpunk 2077 proves that.