And on the reverse note, no: not everything imaginable can be automated. This thing I do monthly, which requires running a query nine times, that takes six hundred values as parameters (values that have nothing in common in the database, so they have to be individually listed in the query, I mean, there's nothing distinguishing this list from all the other values it might contain instead), that involves four SQL partitions, five CTEs, eight temp tables, and an hour of pivoting in Excel...no, I can't just "automate" that to have it appear "on a dashboard somewhere" in an application I don't have rights to. Even if the users could agree on the base question of what data they actually want returned (they can't agree, and it's been different every month)!
eta: oh, and this is fun: they all want the "high level numbers" but a minority want to be able to drill down from there to the detail...the detail is between two and six million rows of data. What, precisely, are you going to do with six million rows of data? Read them all? I think that would take an entire human lifespan to do. Also it wouldn't be informative because the question they're interested in is why the events listed in the data occurred, which information is not contained within the data itself. The database (and I) can only say that Event 1209032 occurred, we cannot divine why it occurred instead of a different event not in the data, or what would have happened if it had occurred in a different manner.
I think some people heard the word "Oracle" in connection to a database and thought the things are actual oracles.