The interplay between labor arbitrage and consistent service delivery in IT outsourcing continues to drive business process standardization. New emerging standards, like ITIL, is defining an industry-wide taxonomy for IT service management. These standards are often at a high level and require substantial investment from service providers to define and implement these standards across the various low-level processes they offer. This paper presents a process management system, called Cyano, that uses social networks and recommendation to greatly increase the effectiveness of process capture and knowledge maintenance. We particularly focus on Cyano's social recommendation engine, called SCOOP, which utilizes the intrinsic graph property of process content for recommendation. More specifically, SCOOP maintains a user-process interaction graph Gand computes the user-to-user similarity scores using the random walk with restart on G. Finally, we evaluate SCOOP in the context of a large-scale deployment of Cyano, with thousands of processes and users.