For pervasive computing ideas to reach, and be used, by the general public, they should fulfil an unmet need. Automation of manual tasks in the home, car or at work provides a rich environment for new services and applications. Pervasive computing can be the infrastructure upon which such services can be built. However, without intelligent systems to support pervasive computing environments, users will be overwhelmed by the complexity of the systems with which they are asked to interact. Experience shows that, to be adopted, technologies must address a need without making unreasonable demands of the user, and without imposing huge maintenance and access costs. Intelligent systems will play a major part in enabling the applications that will create user need in pervasive computing environments — facilitating user access to the environment and reducing management cost to supportable levels. This paper will explore the social and technological research issues relevant to the use of intelligent systems in pervasive computing. These include privacy and data protection, user interface and human interaction with intelligent systems, knowledge acquisition, and the management of large-scale systems capable of autonomous action.