This paper discusses courseware authoring for CALAT, a Web-based intelligent CAI system, mainly from the standpoints of courseware design methods and authoring tools, then presents the results of application to courseware in the telecommunications service field. A feature of an intelligent CAI system is its ability to adapt to individual learners, dynamically selecting the material to be presented based on each learner’s comprehension and other parameters. Conventional intelligent CAI systems, however, required nearly the same effort and know-how for courseware authoring as for system development, making it difficult for specialists in the field of learning to create courseware. This paper presents a design method based on courseware knowledge structured around learning targets and a scenario, and shows how courseware with a high degree of adaptation to the individual can be created readily by people whose specialty is the courseware field rather than programming. It further describes tools supporting efficient courseware design and for automatically generating courseware. This system has been used to create more than 300 courseware packages in the telecommunications services field, which have better learning effectiveness than conventional CAI systems.