Currently, if LabelBOT does not find any suggested labels, the pending annotation is deleted again. This could be frustrating for complicated (video) annotations because the user has to draw them again. Instead, we could show the LabelBOT popup that only contains the (focussed) typeahead so the user can choose a label manually. The popup could also contain the explanation text that LabelBOT was not able to suggest any label. Only if the user dismisses this popup, the annotation will be deleted.