Making the best use of knowledge as one of the most important organizational resources is a major aim of sound management practice, striving to achieve reduction of risk, duplication of effort, and uncertainty in the day-to-day work of an organization (Galbraith, 1994; Sanchez & Heene, 1997). A survey of literature on organizational requirements for knowledge management indicates that the matter is also of some theoretical importance, having brought to light several reasons for the attention that knowledge management is currently receiving.

Knowledge sharing, learning from experience, and open access to information about the ways an organization works are made possible by advanced information and communications technologies (ICT). ICT have increased access to and decentralization of informational resources and thus stimulated organizational learning (Walsh & Ungson, 1991). Furthermore, the emergence of the global business community through ecommerce and ebusiness has brought to light considerable benefits that enhanced conditions for partnership and trust can bring to distributed organizations. In particular, their access to global markets is considered to be the key to their success, as it enables timely and effective response to changing customer demands (Nonaka, 1991).

The concepts of knowledge management and organizational learning have also been related to the particular, situated aspects of day-to-day working practices of a distributed organization. They are applied to the study of teamwork in physical, media, and virtual environments, where cognitive and social aspects of organizational processes determine the requirements for technological platforms and applications (Davenport et al., 1998). Studies of individual workers’ circumstances, in the workplace, at home, and on the move, emphasize their need to access an organization’s knowledge networks, through both formal and informal channels (Baker, 1994).

In this context, there is a clear distinction between “hard” and “soft” knowledge. Hard knowledge is treated as some kind of commodity that can be exchanged, transferred, and coded into documents, records, charts, and other similar informational artefacts (Teece, 1998). Increasingly, however, we consider the importance of “soft” knowledge, the implicit communication and tacit understanding of the particular situations in which knowledge is put to use. Tacit knowledge can be transmitted through concrete examples, experience, and practice through personal interactions (Boisot et al., 1997).

The work presented in this article is based on the study of tacit knowledge that is communicated within and between teams of knowledge workers “on the move.” Their experiential knowledge is shared through personal interactions in real-life tasks, carried out by agile teams in construction who work together on a joint project while being employed by different subcontracted firms. Construction teams are agile in the sense that they have flexible, ad hoc organizational structures, which increase their mobility and their responsiveness to continuously changing circumstances on the construction site. Although they have strict formal channels of communication and established procedures and protocols, most of their day-to-day working life is oriented toward informal social networks created through experience. It is those informal networks that help individuals cope with the continual change that characterizes their work environment, and enables them to work as part of distributed teams, across physical, cultural, and organizational divides. For these reasons, agile teams in construction are taken as an example of a distributed organization held together by informal social networks that support shared tacit knowledge of the ways the teams can work together effectively.

The study builds on recent research within the academic community concerned with the study of distributed organizations, but focuses on a practical knowledge management project, CICC.1 The main aim of the project was to improve relationships between people, technology, and information content in the real-world environment of a large construction project. CICC technologies were designed and implemented bearing in mind that introducing them into a real-life working environment requires sensitivity, awareness, and understanding of the human dimensions of construction teams’ work.

As teamwork is essentially social in nature, the CICC technologies predominantly address the need for increased connectivity between people in order to improve knowledge sharing within and between construction teams. Knowledge management applications developed within CICC include intranets, organizational memories, and general knowledge management tools. Intranets enable networked access to people and the informational networks to which they belong (Rosenberg, 2000). Knowledge repositories and organizational memories provide access to the pool of knowledge that an organization has about its past projects (Perry et al., 1999). Management tools are focused on knowledge networking practices (Rosenberg & Holden, 2000), recognizing not only the strategic importance of data resources that underpin the creation of knowledge assets, but also the need to support exploratory learning of individuals in organizations that “will enhance their personal effectiveness and their overall capacity to interpret and make sense of experiences” (Boisot, 1999: 266).

COORDINATING AGILE TEAMS

Agile teams working on a shared problem face challenges at two levels. First, the usual composition of such teams requires that participants have diverse expertise and thus their distinct contributions to the shared task have to be coordinated. Second, agile teams are created to work together on a specific project and are dissolved when the project is finished. Thus, the team members rarely know each other well prior to starting work on a joint task, and consequently their individual perspectives on the collaborative situation must be coordinated “on the fly.”

In the study of coordination of work within agile teams, the focus was on the effect that these two factors—diverse expertise and team agility— have on the coordination of joint activities. The main aim was to find out how various informational resources available in the work environment aid (or obstruct, as the case may be) the coordination of both taskoriented and communication-oriented aspects of collaboration. Particular attention was paid to the role of advanced communications technology in facilitating such coordination when the teams are not only agile, but also geographically dispersed. New technologies such as multimedia, mobile telephones, wearable computers, and video have the potential for supporting working environments characterized by professional and cultural diversity.

It is important that this technology fits in unobtrusively with the working practices of particular communities, supporting the interactions and activities that have naturally evolved as part of social life in the workplace. (Winograd & Flores, 1986)

Coordination of distributed work activities is crucial for the success of a collaborative task. Such coordination is also central to the study of collaboration within CSCW (computer-supported cooperative work) where the focus is on “how co-ordination emerges, how it is maintained and what happens when it breaks down” (Rogers, 1993: 295). In particular, the entire working environment is examined to find the features that facilitate coordination of distributed activities. The most important of these are the shared artifacts, also referred to as common artifacts (Robinson, 1993), boundary objects, communicative artifacts, collective cognitive artifacts, or common objects, that are used by teams to create and maintain mutual understanding of the work process.

Shared artifacts provide the resources for people to focus their collaborative activities and to obtain facilitative feedback from each other about the current state of the activity. They can be regarded as loci, or physical manifestations of what is essentially purposeful social action. For example, a simple common object such as a set of pigeonholes in a university department has a structure consisting of slots for holding mail, documents, and messages for individual staff. It also has an informative function, since it makes it easy to see if an individual has any mail. A pigeonhole that is full up may “mean” that the owner is not back from a vacation or conference, or perhaps that there is substantial marking to be done. The set of pigeonholes also shows at a glance the size of the department, thus providing some information about both individuals and their organization.

The starting assumptions behind the work presented in this article are that shared artifacts are the key to investigating the coordination of distributed activities, and that this coordination takes place at two levels simultaneously. At the task level, participants coordinate their perspectives on the job in hand in order to establish the common ground for joint problem solving. At the communicative level, participants negotiate in order to reach agreements on how what is said and done should be interpreted. The distributed activities are thus coordinated by participants monitoring the progress, noting the changes, and providing feedback about their own actions and reactions, relying on the shared artifacts for this purpose.

Personal interactions follow a general pattern: the participants establish contact, make their situation visible to others, and then together build a shared environment where they cooperate to solve the current problem (cf. Gumpertz & Hymes, 1972). Participants focus on shared artifacts in the process of negotiating the meanings of words or images presented there (Robinson, 1993). They thus create the common ground: “a sine qua non for everything we do with others … the sum of [the participants’] mutual, common or joint knowledge, beliefs, and suppositions” (Clark, 1996: 92).

Within the boundaries of the common ground, the participants can identify the objects referred to, come to understand each other’s goals and purposes, cooperate, and coordinate their actions. Indeed, common ground is regarded as fundamental to all coordination activities and to collaboration (Clark & Brennan, 1991). One of the key research questions for the CICC project was how people create the common ground in situations where the contact between them is influenced or mediated by technology (Rogers, 1993; Hindmarsh et al., 1998; Dourish & Bellotti, 1992).

Assuming that common ground is fundamental, it is important for the designers of technology to understand what representations should be employed in order to design usable interactive technology. In her analysis of technology-based shared artifacts, Rogers (1993: 296) distinguishes between two main kinds of mediating mechanisms:

explicit representations which are intended to provide specific information about the status and properties of actions and artifacts that constitute the work, and implicit representations where an action or some form of communication signifies a change in status of an object or process.

Some of the mediating mechanisms have universally accepted meaning, while the meaning of others is negotiated and interpreted anew in each particular context. A further dimension to the mediating mechanisms is added in the practice theories of culture that emphasize the importance of language use in “cultural meaning making.” Placing language at the center of the creation of social meaning and the consequent building of communities, practice approaches also take into account the resources available in the community’s environment. The resources that enable participants to create distinct communities include various designed artifacts, such as documents, instruments, computer applications, and others, that are at the center of personal interactions among team members. In such situations, the technology provides the shared artifact needed for the coordination of distributed activities, and the nature of computerized representations, both explicit and implicit, becomes a critical issue. Such technology is expected to facilitate not only access to the data that an organization has, but also to facilitate the processes that shape human cognition and communication in the significant social and cultural contexts, thus fitting in with normal human activities in the workplace.

THE PEOPLE AND INFORMATION FINDER

The working assumption behind the design of usable interactive technology within the CICC project was that it forms an integral part of the entire information environment created by the interaction of people, organizations, and artifacts, where information is generated, exchanged, stored, processed, internalized, and externalized. An interactive multimedia technology prototype, People and Information Finder (PIF2) was designed to provide explicit representations of people, their organizations, and teams, as well as the projects on which they were working. Multimedia facilities were used to increase the richness of the information about the PIF people and their workplace. These included visual channels with both abstracted and image-based data. Such increased media richness was expected to enhance the creation of common ground among the PIF users, thus enabling tacit knowledge sharing in communication mediated by the PIF. The main research questions concerned the content of multimedia displays and explicit representations, as well as their informativeness and implicit representations.

As the main aim was to improve communications and to offer a richer information environment for the PIF people, a key research issue was to find out how the explicit and implicit representations in the PIF could help them to initiate and maintain contact with one another. Such a facility is particularly useful in situations where a stable organizational form is absent, as a team is created for the purposes of a particular project or a task and is dissolved once this is completed. The work on the task itself represents a period during which cooperation is vital. Problems of lack of shared culture have to be addressed and this is usually done with a series of induction meetings and seminars at the start of the project. This is a vital yet time-consuming team-building stage, and is undermined by individuals joining projects at later stages with consequent integration problems (Rosenberg et al., 1997). Poor communication causes serious problems in day-to-day activities that require continuous cooperation and coordination (Fruchter, 1998).

One of the key roles intended for the PIF was to become an important shared artifact and a vehicle for soft knowledge management practices, designed to help those who need to learn about a given project and its organization. Such learning would take place in “communication at arm’s length” mediated by the PIF, which would enable people to assimilate the project culture without disturbing the established flow of activities on the site. The social relevance of the PIF as the shared artifact could then be analyzed as a matter of system-in-use that systematically interacts with the organization that uses it.

As a technological artifact, the PIF is a web page populated with information about its owners, using standard web technology and programming in html underpinned by Java script. It comprises a technological platform; more precisely, a configuration of communications technologies such as telephone, network technology such as the web, and advanced interactive technology, such as virtual and augmented reality. It thus provides an integrated service that helps the user to choose between various channels of communication, from voice link to videoconferencing, and to browse through a collection of similar pages created by different owners according to a predesigned template.

It also provides a richer experience of people and their work environment in a particular organization; in other words, it supports the creation of common ground between the PIF people. They obtain information from people, from informational resources such as documents or databases, or from the real world where PIF owners live and work, and the structure of the PIF page reflects this. It is divided into three main components. The first contains information that helps visitors to locate the owners in their physical space, giving name, address, phone number, and email address, as well as photographs and video images showing their offices, desks, and terminals. This makes it possible for the visitor to choose how to contact them, via telephone, ordinary mail, email, or videoconferencing. The second, “nearest neighbors,” provides information about the organizational space that a team occupies, giving similar information about accessing other PIF owners who are engaged in comparable or related work. The third component describes the owners’ projects and activities, both present and past, comprising the part of the real world in which the owner’s work is done.

As a knowledge management tool, the PIF provides a facility for accessing information about an organization or a team in the construction industry, the people who work for it, and the activities in which they engage. The status of information ranges from personally owned (or private), which is accessible only to the owner, to information that is distributed (or shared) between members of a group. Alternatively, information may be public (or visible), which the owner is willing to show to the general public or to a selected group of collaborators. This nonproprietary information forms part of the informational resources used by the organization or the group as a whole.3

The PIF thus has the potential for enabling innovation in cooperative work within agile teams. In order to understand more about this potential, a preliminary study of communication mediated by the PIF prototype was carried out in interviews with a small group of informants. Their responses were elicited in open and in structured interviews, as well as focus groups. The data analysis was oriented toward explicating the informational links between what visitors could observe in the communicative situations displayed on the pages and how they interpreted the information presented there.

The main aim was to find out what general interactive strategies people use when learning about a team, its people, and their work. In particular, the focus was on discovering what knowledge, presuppositions, and beliefs people bring with them to joint activities, and how the external representations designed in the PIF could influence the creation of common ground.

ANALYZING SOCIAL KNOWLEDGE

In the analytical framework developed for studying the use of the PIF in a real-life setting, the Common Ground Framework was extended and adapted for the study of representations in the PIF pages in order to identify the features that would facilitate the communication it mediated.

The original framework views communication as joint activity between two or more participants engaged in a shared task (Clark & Brennan, 1991; Grosz & Sidner, 1986; Clark & Schaefer, 1989) who, in the process of carrying out this task, have to establish the common ground for mutual understanding and trust. Clark (1996: 43) divides the common ground into three parts:

  1. Initial common ground. This is the set of background facts, assumptions, and beliefs the participants presupposed when they entered the joint activity.

  2. Current state of the joint activity. This is what the participants presuppose to be the state of the activity at the moment.

  3. Public events so far. These are the events the participants presuppose have occurred in public leading up to the current state.

Initial common ground. This is the set of background facts, assumptions, and beliefs the participants presupposed when they entered the joint activity.

Current state of the joint activity. This is what the participants presuppose to be the state of the activity at the moment.

Public events so far. These are the events the participants presuppose have occurred in public leading up to the current state.

In the modified framework, the presuppositions determining the initial common ground were classified into those related to the background knowledge characterizing participants’ expertise and the more personal, experiential knowledge of teamwork on construction projects. The former includes, for example, individual expertise in architecture, construction management, structural engineering, or any other relevant profession that determines the nature of their contribution to the process and task activities. The latter is based on the experience of working on construction projects, with particular partners, and the trust that such an experience generates. These form the starting context for joint activities and influence the manner in which participants recognize standard procedures, identify their own roles and responsibilities as well as those of others, and decide how to organize joint activities. Special attention was given to the social knowledge that forms part of the initial common ground and, more specifically, how the explicit representations of the PIF interact with the creation of the tacit knowledge about its owners.

The visitors to the PIF pages were observed in a “conversation at arm’s length” with the owners, where the main goal was to learn about the “project people.” Observations of visitors navigating through the sample were focused on ease of use and on the extent of the support for learning that it offered. Two main research questions guided both the data collection and the interpretation of the informants’ responses. The first was concerned with the informativeness of various pages in terms of content. The second question was whether or not the principles that govern social action and face-to-face interaction in the real world would be equally valid in the mediated interaction in the media space and the virtual world.

REPRESENTATIONS OF THE PHYSICAL AND SOCIAL SPACES

The information that the visitors picked up from the web pages was considerably richer than that explicitly expressed. Much of this richness was related to the social and organizational characteristics of the owners’ physical environment. For example, the “god’s eye view” of the office layout helped them to infer the social structure of the group inhabiting the space. The assumptions normally made about the social significance of physical spaces in the real world were also made about the virtual world.

Knowledge of the ways in which real organizations structure their working space helped visitors to interpret the social aspects of the situation presented on the screen. For example, they understood that people who have their own offices are higher up in the organizational hierarchy, and that those who are physically colocated usually belong to the same working group. The external representation of the office space thus plays an important part in identifying the owners’ status and relationships.

REPRESENTATIONS OF PERSONAL SPACES

Interpretations of individual web pages confirmed the importance of links between the images of the physical environment and the socially relevant information inferred from it. Different views of the physical location reflect the personal image of the owners as if they were wearing “electronically augmented clothes.” For example, the scene of papers scattered all over the desk makes the owner appear as an untidy but creative person, someone who works with paper and computers, which indicates a high level of education and might be seen as “serious, reliable and responsible.”

A detailed view of the workspace also contributes to familiarization with the owner. The owner of a particular space may be assumed to be creative and clever, as they work with computers and have lots of paper scattered around, working on knowledge-intensive tasks, spending much of their time at the terminal if there is a comfortable chair in front of the computer.

THE INFORMATION ENVIRONMENT

Originally, the design intention was for the personal page to provide information about the owners’ availability by making their situation at work more visible to potential visitors. The possibility of intrusion into the owner’s space at an inopportune time would thus be reduced. A camera placed in the owner’s office was the main source of this kind of information, enabling visitors to make informed decisions about whether the owners should be approached via an intrusive medium such as telephone, or videoconferencing, or whether an email message would be more appropriate.

Messages left on the screen when there is no activity are a very important guide to availability. For example, Post-it® notes saying “Back this afternoon/tomorrow/next week” give immediate guidance on when the owner can be contacted. A collection of such messages creates the context for interpreting similar situations when there is no action on the screen. Thus, the view of an empty office without a message would mean that the owner is momentarily absent, perhaps talking to a colleague in the corridor or having lunch, while a message saying “Goodbye” would mean that a longer, or maybe indefinite, period of absence was intended.

People who share the working environment with the owner can provide an alternative point of contact if the owner is not available. Particularly useful features are the links with closest collaborators within the intellectual space they share, indicating to the visitor that close collaboration is no longer restricted to people inhabiting the same physical working space. The PIF is therefore an informational resource that can transcend the limitations of the real world; this is where the visitors saw possible new uses of this technology in their own lives. The general image was of an impressive technological achievement providing a novel service.

The overriding principle in visitors deciding who to contact was the awareness that they may be intruding into other people’s working spaces and they were therefore looking for signs of invitation. In general, they expected a virtual host to help them orient themselves in the organizational environment and to act as a focal point to which they would return when they get lost in the virtual space they were navigating. The host was also expected to sanction the appropriate forms of contact and determine the acceptable degree of intrusion in any given circumstances.

Particularly important was the realization that the owners had full control over the cameras in their offices, so when they did not wish to be watched they could turn the camera off. This meant that if the camera was on, the visitors were invited to enter, but they were still aware of different degrees of intrusion possible in the circumstances. For example, accessing a database or clicking on a document was equivalent to a permission to “enter without knocking,” sending an email message was comparable to leaving a message on the door, but videoglance helped with the choice whether to knock or not. Videoconferencing, being the most intrusive call for contact, required the most explicit permission.

The scenes offered by the camera were interpreted in the light of such social considerations. The “degree of closeness” between owners and visitors was often determined by interpretation of symbols: for example, a highlighted name in the “nearest neighbors” section was seen to be an invitation to contact that person rather than somebody else. Rules that are observed in face-to-face interaction also played a part in decisions to establish contact, so that people whose faces were turned toward the camera were generally considered to be more friendly and approachable.

Most visitors saw the intranet page as providing much more than a home page on the web. They used it as a traveling map helping them to locate the owners within the wider context of their organization. The context was referred to simultaneously as the physical and informational space the owners occupied. The global function of the page from the visitors’ point of view was that of an augmented map enabling a view of individuals in relation to their neighbors and to other sources of information available in the real and virtual worlds.

CONVERSATIONS IN PERSONAL SPACES

Conversations among construction workers are normally centered on the object being designed or under construction and the actions required to get the design or construction process going. In this context, people suggest new tasks, justify their actions, and check that the tasks have been done. At the same time, they negotiate their responsibilities, evaluate contributions, and establish individual authority.

The content of their conversations also refers to actions on design or media objects, such as showing a file or drawing, creating the design objects or media objects, such as PowerPoint slides that represent the design objects. In the course of a team session, design decisions have to be made, both choosing among the alternative solutions to design problems and about the responsibilities and contributions of individual team members. In the traditionally organized physical workspace, these conversations have naturally evolved. The powerful new technology promises that such naturalness can be maintained in conversations where people are not physically present together, but are in contact through telephones, videoconferencing, email, virtual reality, and other similar technology applications.

The findings of this study suggest that more work needs to be done before this promise is fully realized in practice. They also suggest that a significant aspect of future work should be devoted to theoretical issues of interdisciplinary research. A main problem that needs to be addressed at the start involves the study of interactions in physical, media, and virtual spaces, in particular the linguistic approaches to conversation analysis. These are traditionally focused on the study of face-to-face conversations where two participants jointly create the “frame” of the conversation. The findings of this study show that the initial common ground is much more varied and comprises considerably more social knowledge and assumptions than the traditional studies within the common ground framework have so far acknowledged.

Furthermore, in a real-life workplace there is a greater number and a greater variety of conversations than face-to-face encounters. It is often impossible for people to participate actively in a conversation, so they may overhear, monitor, or ignore it, while still being aware that a conversation has taken place. The understanding of the overhearers is different from the understanding of the active participants (Clark, 1996), and this should be taken into account, both in scientific coverage of interactions in the workplace and in the design of technology that supports them.

Many joint activities of teams are carried out “at arm’s length,” instead of face to face, where the possibilities of misunderstanding are greater and the facilities for repair reduced. It is therefore important for communications technologies to observe the established representations of information in terms of rules, regulations, and etiquette that is easily accepted by team members. It is even more important for the technologies to enhance the representations of information to include more personal and less structured requirements, such as helping people to build the common ground that ultimately leads to trust.

CONCLUSIONS

Several principles governing social interaction in establishing face-toface contact also apply to the situation of virtually visiting an organization and its people using the intranet example described in this study. The prototype makes a person’s environment visible to others at remote locations, and helps in creating a common ground to underpin joint activities. An important social (and design) principle is related to intrusion into another person’s space, be it physical, organizational, or informational. Invitations into such spaces can be communicated indirectly and visitors will look for them in the objects presented on the page, such as the videoglance, shrunken screen, and highlighted names. This will happen even if the designers of the page have not explicitly intended for these objects to have such a communicative function.

Another important principle concerns the construction of initial common ground. Visitors’ interpretations of the information presented on a personal page will be considerably richer and with more social detail than the literal meaning of the text, pictures, or graphics presented. This is where the metaphor of “electronically augmented clothes” is particularly apt. Decisions are made not only about the appropriate means to obtain information by accessing the sources that it makes available, but also about personalities and possible relationships between them, the nature of the organization, and maybe even the quality of work that can be expected of them.

The information displayed on the intranet pages is therefore interpreted within a rich context of natural human communication, where the accepted social norms that determine appropriate behavior apply. The extended common ground framework developed here provides the conceptual basis, as well as methods and techniques for discovering and formulating the social constraints that regulate joint activities and communication at arm’s length. It is the framework for the design of technology that brings together people, their social relationships, and the resources they need in order to carry out the joint activities in the workplace, which are the stated aims of knowledge management in practice.

NOTES

  1. EU AC017, CICC (Collaborative Integrated Communications for Construction). A detailed report is available on the author’s website, http://www.rhbnc.ac.uk/∼uhtm059/index.html.

  2. The People and Information Finder is a multimedia prototype developed as part of the CICC project.

  3. The team involved in the development of the PIF prototype consisted of Nicholas Farrow, David Leevers, Mark Perry, and Duska Rosenberg (cf. Rosenberg et al., 1997).