In particular, I enjoy getting to know the people who work at a local establishment and understand their taste. For instance, when it came to fine teas, I got to know one tea merchant well enough that she'd say, "This new green came in and I hated it! I thought you'd love it!" because we had entirely opposite tastes when it came to the astringent characteristics of tea. (Think a Pi Lo Chun versus a Sencha, for instance.)
While we are getting better at recommendations, the best salespeople still have a better understanding of what makes you like a product than "people like you also purchased..." algorithms. This may be particularly important in markets that are about different things for different consumers: wine selection may be about taste, or it may be about social signaling. Same is true in the transaction: it may be about the best price, building a relationship, having a relaxing escape (buying cigars, for instance), or any number of other things.
Even when salespeople don't have a better understanding of your tastes they're more likely to be persuasive and more likely to add the perceived value to it... a salesperson who admits to hating something she recommends is a more memorable and compelling call to action than an algorithm
I'm sure Amazon's algorithms often do have a better idea of what I'd enjoy than many of my friends, but friends' recommendations are still more likely to induce me to act than a personalised promo email.
While we are getting better at recommendations, the best salespeople still have a better understanding of what makes you like a product than "people like you also purchased..." algorithms. This may be particularly important in markets that are about different things for different consumers: wine selection may be about taste, or it may be about social signaling. Same is true in the transaction: it may be about the best price, building a relationship, having a relaxing escape (buying cigars, for instance), or any number of other things.