Book recommendation systems are increasingly common, from Amazon to public library interfaces. However, in the land of archives and special collections, such automated assistance has been rare. This is partly due to the the complexity of descriptions -- EADs describing whole collections -- and partly due to the complexity of the collections themselves -- what is this collection “about”, and how is it related to another collection? The American Philosophical Society Library is using circulation data collected through Aeon to automate recommendations. In our system, recommendations are offered in two ways: based on interests (“You’re interested in X, other people interested in X looked at these collections”) and on specific requests (“You’ve looked at finding aid Y, other people who looked at finding aid Y also looked that these finding aids”). We will discuss the development of this system, the central role of patron privacy, possibilities for generalizing the project for other environments, and future plans. We will also discuss ongoing concerns and issues. For example, do these recommendations increase the use of already highly used collections at the expense of less well-known resources? What nuance are we missing by focusing on the collection-level data, and what alternatives could be developed?