Allow NLU evaluations to be case insensitive
Because it is case sensitive, the automated NLU tester almost always fails when testing values which may be contained in an Alexa built-in slot.
It also appears random as to whether the values in the slot are uppercase, lowercase, or even mixed between two words.
Example: For the slot AMAZON.GB_CITY, it matches the slot values, but shows a fail because sometimes they are lowercase.
Of course, if I submit all lowercase, then it fails on the ones which are uppercase.
It would be very easy to make the nlu tester ignore case sensitivity, which would make it more useful.
E: Ampthill, A: ampthill FAIL
E: Crowthorne, A: crowthorne FAIL
E: Newbury, A: Newbury: PASS
E: Chesham, A: chesham FAIL
Jonathan H shared this idea