Introduction of the inventory gives the Sense algorithm the possibility of exploiting the list to aid and match detections (that, I believe, was the intention) but there are a couple of things (at least) making this a non-trivial task:
-
There is no way to know if any given user-entered list is accurate or complete.
-
If a potential match is found, the level of confidence for the match needs to come after the “native” detection and confidence. That said the list can inform the algorithm but that is certainly a non-trivial task and presents a certain catch-22 in implementation.
I have confidence that Sense has this fairly high on their list of priorities because it could clearly improve detection and user experience. Not easy though.
This thread (and others) hint at other issues: