I don't know what people who do this for a living get paid for doing, but whatever it is, it probably is not enough. I've encountered, as I'm sure you have too, book indexes that are simply worse than useless. (The only thing I can think of worse is a book that's not indexed at all.) There's nothing more frustrating to someone seeking information between two covers of a book who cannot do it because a good tool has not been provided to help him. That's what the index is. There's no sense in having one at all if it's crappy. The only kind acceptable is one that's complete and accurate. I've been working on building one now for several days. And I've got several days more to go.
Some observations. First, it is damned tedious and hard work doing this. You must comb every page of your galley proofs to find items that should appear in the index. You're constantly having to make decisions. Does this belong in the index? Does it have to be cross-referenced? Which category identifies it? Are there other words or phrases that should be indexed under this entry? And so on and on. Nobody gives much thought to the fact that somebody has to make these decisions, and the quality of the finished index is directly related to the quality of countless decisions like this. Second, electronic indexing until we have real AI is never going to be in the same league and good ole human brainpower. Finally, thankfully we have some electronic tools that ease the burden somewhat. The page proofs I'm working with are both hard copy and an electronic .pdf file. A free .pdf file reader called Foxit will search the entire manuscript for certain words or phrases and give you a list. So at least you don't have to work entirely with hard copy, like I had to do years ago with book on Stephens.
But this is a hell of a lot of work no matter how you do it.