IT is increasingly automating business processes, but these are formalized processes that have limited adaptability . While the Semantic E-business vision can enhance business processes and business information exchange, it is people who conduct the non-routine, value-added work that cannot be automated. Most knowledge work is collaborative, informal, and situationally adaptive . People use flexible technologies, such as document editing tools, email, chat, and shared repositories that are not well integrated with formal processes. We call this kind of work "business activity."
As an example, consider Company A, which outsources much of its business by issuing requests-for-proposals (RFPs). It has an automated process for managing the flow of RFPs and proposals. Company B responds to many of these proposals using the automated process. But the work Company B does in responding to the RFPs is a complex business activity too variable to be automated. The responding-to-RFPs activity cuts across business functions; different teams must be assembled; people switch roles and move on and off teams; coordination, tracking, and decision making under time pressure are critical. Company B also needs to evolve best practices to effectively reuse content and processes from its experiences. Such business activities require support that does not inhibit its informal and locally adaptive nature. We call this support Unified Activity Management (UAM) .
The key abstraction is a generic notion of human collaborative activity, and our hypothesis is that an explicit shared representation of Activity is needed. An Activity description articulates the actors (people) and roles involved, the resources used (tools, artifacts, people), the results produced, the events it is bounded by, and its relationships to other Activities (such as sub-Activities or dependent Activities). All the people involved can see the Activity descriptionsthe metaphor is a shared checklistand they can modify and extend the descriptions. Figure 1 shows a shared checklist for the respond-to-RFP activity example (RFP #0518). The main sub-Activities (assess, assemble, and so on) come from an Activity Pattern representing the major phases of Company B's practice for responding. Different people, resources, and events are associated with each of the Activities and sub-Activities. The people involved alter the sub-Activities and add sub-sub-Activities as needed to help coordinate the team's response to this specific case.
The main objectives of this shared Activity representation are:
The Semantic E-business vision should include support for informal human-driven business activities by providing a unified way to represent business activities, enabling different systems to interoperate, enabling the evolution of business practices, and integrating with formally modeled business processes.
The philosophy of UAM is to rely on social interaction mechanisms to control activities and to impose only as much restraint as is needed. Thus, UAM is a natural complement to formal workflow and B2B processes.
We think of Activity descriptions as metadatathe glue tying together system resources around the generic semantics of activity. Thus we encode Activity descriptions in RDF. The Semantic Web is an ideal platform to provide the flexibility, extensibility, and data integration necessary to support the inherent variability and adaptability of business activities.
We are developing an OWL ontology of Unified Activity. The core ontology captures the essential features and structure of the generic concept of collaborative activity, which enables activity-like information from different applications and contexts to be unified, exchanged across systems and companies, and presented in a consistent vocabulary to people. The core ontology is extensible to accommodate the specific features of different business activity domains (analogous to the analysis of processes in ). Thus, activity-like entities in different business domains (for example, tasks in a workflow) can be exchanged and understood at the level of the core activity semantics.
The default mode of UAM is to support flexible and open interactions between people. Some interactions need more structure; and we are exploring extensions, such as access control policies and constraint handling, in order to interoperate with formal processes, such as project planning and workflows, without turning UAM into another workflow language. The philosophy of UAM is to rely on social interaction mechanisms to control activities and to impose only as much constraint as is needed. Thus, UAM is a natural complement to formal workflow and B2B processes. A workflow does not need to micro-manage people; it can delegate complex social activity to people through UAM, which can help them coordinate their work, collect the relevant materials, pass the results back to the workflow, and produce an audit trail of their work to satisfy compliance requirements.
Figure 2 summarizes the positioning of UAM. The Unified Activity representation is managed in an RDF-based Activity Metadata Repository that integrates information from various sources. UAM supports people by providing activity contexts for their informal collaborations and by being an intermediary to formal workflow processes. UAM provides a semantic paradigm for effectively extending e-business systems with the intelligence, adaptability, and creativity of knowledgeable and experienced workers.
Our goal is to evolve a Unified Activity ontology to become a semantic standard that will enable systems at different organizations to interoperate through the semantics of Unified Activity.
We are exploring this paradigm for IBM Workplace, a platform for building business applications integrated with collaborative components. An early example of an activity-based capability in Workplace is Activity Explorer , which allows people to easily share a variety of collaborative objects and to aggregate them into activity threads. We are currently putting Unified Activity metadata over these threads, enabling people to create and evolve Activity Patterns. We are also developing exploratory clients, both Eclipse-based rich clients and Web clients, to validate Unified Activity concepts and to study how people actually use them.
Our goal is to evolve a Unified Activity ontology to become a semantic standard that will enable systems at different organizations to interoperate through the semantics of Unified Activity. Our vision is to leverage the semantically rich OWL-S Web Service descriptions to make Activity Patterns discoverable by OWL-based B2B processes. Consider, for example, an exception condition in a supplier's B2B process that requires a renegotiation of the customer's specifications. Both the supplier and the customer can have Activity Patterns for engaging in such renegotiations (each customer will have its own patterns, which will change over time). The supplier process can discover these and trigger the supplier's and the customer's respective UAM systems to instantiate joint renegotiation Activities involving the appropriate people and including the relevant documents, data, and deadlines. Because the supplier's process has access to these Activities, it can monitor the progress of ongoing renegotiations. Further, the supplier has a rich store of detailed renegotiation activity descriptions that it can analyze for ways to improve its B2B process.
We believe that effectively incorporating semantically based support for collaborative business activities qualitatively enhances the adaptability of the Semantic E-business vision.
4. Muller, M., Geyer, W., Brownholtz, B. Wilcox, E., and Millen, D. One hundred days in an activity-centric collaboration environment based on shared objects. In Proceedings of CHI 2004. ACM Press, NY.
©2005 ACM 0001-0782/05/1200 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2005 ACM, Inc.