From Office Automation to Intelligent Workflow Systems

• Authors: Dirk E. Mahling;Noël Craven;W. Bruce Croft
• Venue: IEEE Expert: Intelligent Systems and Their Applications
• Year: 1995

The authors trace the technological and conceptual developments of their approach to the support of task management over a period of more than ten years.Supporting teams, groups, and whole organizations in routine and exceptional tasks has been a long-standing goal of information systems. Currently, workflow systems tackle this problem anew. In this article, we present the evolution of task support research in an attempt to contextualize the current efforts in workflow systems research and show how we can benefit from earlier work in office automation.We can describe most endeavors in business or daily life as processes. These processes consist of tasks that usually involve more than a single person or intelligent machine agent. Groups of people and machine agents perform tasks and processes in a wide variety of environments to achieve their objectives and goals. These processes span a wide conceptual spectrum from routine, repetitive procedures to very complex and highly unique projects. The groups and organizational settings are equally diverse, ranging from processing checks in a large bank, to procurement and large-scale civil engineering projects.The origin of information system-based task-support for individuals and groups dates back to the early days of computing machinery. Yet the limited business success of the "paperless office" and other process-based systems in the 1970s and 1980s lessened interest in this area. Activity in groupware research in the 1980s and 1990s shifted to the definition aspects of cooperative work, such as shared information spaces or shared editing. One of the major problems in interactive process support is spatial and temporal distribution. The competence of the cooperating partners may vary widely. Viewing human and machine agents as partners in communication, cooperation, problem solving, and task execution extends the conventional paradigm of computer supported cooperative work and distributed AI.Business process reengineering and activity-based costing reemphasized the importance of the dynamic aspects of processes and tasks. The growing number of workflow systems supporting BPR attests to this. Many recent workflow systems contain sophisticated graphical user interfaces and networked microcomputers. Other systems build on the experience of previous task-support systems. However, the relationship between the office information systems and the current workflow systems remains somewhat unclear. If the new generation of workflow systems is to benefit from past research, we must examine both the pitfalls and successes of prior task support systems.During the last decade, our research on intelligent support of process has incorporated advances in plan-based expert systems, computer-supported cooperative work, socio-cognitive engineering, and network and communication technology. A line of research on goal-based agent systems for the support of processes has emerged as well.