The tidbits on Siri's personality at the beginning of this article are a good reminder of how a good UX can make people very forgiving of high error rates. Another example is that, in recommender systems, explaining why you made the recommendation can often make people more forgiving of bad recommendations and increase overall perception of the quality of the recommendations.  In general, it sometimes is easier to get improvements in perceived relevance (with good design) than actual relevance (with better algorithms or data) for some very hard problems.
Shared publiclyView activity