Human technical support agents spend significant time interacting with customers via various channels of voice, email and chat. There is a massive incentive to automate support with autonomous agents with the goal of reducing manual effort and tune taken for problem resolution. As technical support questions are complex and diverse, building a generic agent capable of solving multiple domains is implausible. In this paper, we describe a scalable conversational framework that automates the process of guided troubleshooting called COBOTS (Cognitive BOts for Technical Support). Our underl ying premise is that scalability in such frameworks can be achieved by control and co-ordination across multiple domain expert bots. These bots co-ordinate to (a) understand user problems from natu ral language queries (b) engage in conversation and (c) provide assistance with troubleshooting. AU of the above is done with mini mum human assistance. COBOTS framework comprises of User Bots that monitor customer infrastructure for issues, the Orchestrat or bot which co-ordinates and controls various request-response pairs and Domain Expert bots which handle issues pertaining to their domains, respectively. In a real environment, we have dep loyed an implementation of our COBOTS framework which can co-ordinate and control user queries across 11 different technical support domains. When evaluated by two different teams of expert support users, it was observed that more than 75% of the time our application was able to provide relevant solutions for their queries.