Companies tend to use algorithms for interviews for new college grads not because that's what the job entails, but because that reflects what the candidate's background was about (college).
If someone is fresh out of a 4 year program that was all about data structures, algorithms, operating systems, etc, then demonstrating mastery of those topics tells an employer that a person is smart and can learn large amounts of complex technical material.
I agree that algo questions are useful for recent grads. I find it more frustrating when I see those questions given to candidates for senior positions with 10 or more years of experience. Keeping fresh on the big-O of a bubble sort and rebalancing a binary tree ends up being in addition to building enterprise software solving real-world problems day to day.
If someone is fresh out of a 4 year program that was all about data structures, algorithms, operating systems, etc, then demonstrating mastery of those topics tells an employer that a person is smart and can learn large amounts of complex technical material.