Most knowledge workers have benefited from a model where we learned the skills, and then, we sold our skilled work. Sometimes to other people that didn't have the skills, and sometimes to an organization that needed an army of skilled workers. We didn't even need a lot of experience. It was enough if "I have read all the books".
This model was based on scarcity. It required the use of limited hours, to go through all the content, which was also scarce, difficult to access. I still remember saving for a text book, or pulling journals from shelves at the library.
Today, we have a different version of this same model. But today, we live in abundance of content and information. "I have read all the books" means something different. It means to take the time to sift through all the content and determine what's important and what's not; and even more valuable, determine what's relevant to a customer.
Today, an AI can answer the first part of the question, and pickup on the concepts that are more often used. The second piece, is not as easy -yet-.
There is another layer though. Making a recommendation to a customer, is more than just finding "the best" alternative. It also implies finding the alternative that the customer will feel comfortable with, that aligns with their values, fits their culture, and that will adopt as their own and will be invested in emotionally.
That, an AI cannot do.