It may not be AGI, but that was never claimed.
I'd argue that "AI" doesn't have to be attempting to achieve AGI to be considered AI. For example, "Computer" players in games are AI. This applies whether you're talking something as advanced as a deep strategy game like the Civ series, or something as basic as the ghosts in Pac-Man, or even an automated Tic-Tac-Toe player.
So...is a path finding algorithm "AI"? I dunno. I could go either way.
But like...I seriously can't come up with a real definition of "AI" that wouldn't end up including things that definitely aren't AI. For example, is AI simply any code that seems like it has agency and makes decisions? If so, then is any sort program that triggers execution based on certain detections "AI"? Certainly not. Is it code that attempts to mimic human behavior, even at a basic level? If so, then is AutoHotKey AI? Nope, definitely not.
Were they that desperate for funding that they had to come up with some AI stuff for an already solved problem?
Especially in places where street data isn't complete or works differently (no street names) than in the western cities where Google Maps etc. are developed.
Grab (https://engineering.grab.com) always has interesting blog posts about these topics like: https://engineering.grab.com/road-localisation-grabmaps
Path finding is one of the classic AI research areas, back when it mostly meant making robots able to interact with their environment. Yes, it's an "already solved problem", in that we already have several good enough algorithms derived from all that AI research; but that doesn't change the fact that this is "AI stuff", and it doesn't prevent new AI approaches from being used (perhaps something based on neural networks as a path finding heuristic, since we have a lot of computing power to spare nowadays).