New technology could bring about a revolution in the way care services are delivered. Intelligent building systems take inputs from sensing devices located around a building and then use this information to control devices connected to a network. Our proposed solution is based upon the fact that the physical and logical unit of an intelligent building is a single room. Each room may contain many sensors and output devices which will be controlled and monitored locally by a node. All the nodes are connected together via a network, thus forming a decentralised architecture that enables building-wide collaboration. We utilise AI techniques to produce a system that is capable of learning how to modify its environment in a manner that is consistent with the occupant's preferences, with only a minimum amount of explicit input from the user or carer. This differs from existing building automation systems, which tend to be pre-programmed by an expert and impose a rigid set of conditions on all of the building's occupants. A behaviour-based approach is used for the internal architecture of the nodes. In our system, each node could be thought of as an agent, with its own set of both fixed and dynamic behaviours. (3 pages)