Interaction technology: Speech act based information technology support for building collaborative...

23
Interaction technology: Speech act based information technology support for building collaborative relationships and trust Kuldeep Kumar * , Irma Becerra-Fernandez 1 Alvah H. Chapman Graduate School of Business, College of Business, Florida International University, 11200 SW 8th Street, Miami, FL 33199, United States Available online 27 September 2005 Abstract While relationships and trust are now commonly accepted as central to conducting business both within and across organizations, literature provides only minimal guidance as to how relationships and trust in inter and intra-organizations are created. Moreover, the role that information technology could play in building trust has largely been ignored. This paper describes an Internet-based process for building trust between collaborative commerce partners. Integrating concepts from Winograd and Flores’ concept of speech act theory based bconversations for actionQ with research on bclosed loop cyclesQ of trust and relationships in the disciplines of management and marketing, the paper first develops a framework for understanding trust and trust-building processes. The framework characterizes the process of building of trust as the management of commitments inherent in speech acts between requesters (customers) and performers. Furthermore, joint sense making during the conversation for action contributes to greater transparency thereby also increasing the levels of mutual trust. The process framework is used to develop requirements for information technology support for a process and tool for building trust. The paper then goes on to describe an implementation of the conversations for action and the closed loop cycles through a web-based software tool based upon Winograd and Flores’ work. Experiences with managing commitments and closing the loop are presented through a case study in an organization that develops and maintains menu-driven voice applications for the call center industry. The case study shows the inadequacy of traditional communication technologies in managing complex, geographically distributed collaborative commitments, and shows how the use of the software tool contributes to a greater level of satisfaction and closing of the performance loop. The paper ends with a reflection on the nature of the tool, its possible uses and misuses, and the role of human wisdom in its use. D 2005 Published by Elsevier B.V. Keywords: Speech act theory; Conversations for action; Building relationships and trust; Interaction technology 1. Introduction ‘‘Fresh flowers are blooming on the battle-scarred landscape where once-bitter rivalries among suppliers, 0167-9236/$ - see front matter D 2005 Published by Elsevier B.V. doi:10.1016/j.dss.2005.05.017 * Corresponding author. Tel.: +1 305 348 3156. E-mail addresses: [email protected] (K. Kumar), [email protected] (I. Becerra-Fernandez). 1 Tel.: +1 305 451 2830. Decision Support Systems 43 (2007) 584 – 606 www.elsevier.com/locate/dss

Transcript of Interaction technology: Speech act based information technology support for building collaborative...

  • The process framework is used to develop requirements for information technology support for a process and tool

    1. Introduction

    Decision Support Systems 43 (* Corresponding author. Tel.: +1 305 348 3156.for building trust. The paper then goes on to describe an implementation of the conversations for action and the closed

    loop cycles through a web-based software tool based upon Winograd and Flores work. Experiences with managing

    commitments and closing the loop are presented through a case study in an organization that develops and maintains

    menu-driven voice applications for the call center industry. The case study shows the inadequacy of traditional

    communication technologies in managing complex, geographically distributed collaborative commitments, and shows

    how the use of the software tool contributes to a greater level of satisfaction and closing of the performance loop.

    The paper ends with a reflection on the nature of the tool, its possible uses and misuses, and the role of human wisdom

    in its use.

    D 2005 Published by Elsevier B.V.

    Keywords: Speech act theory; Conversations for action; Building relationships and trust; Interaction technologyInteraction technology: Speech act based information technology

    support for building collaborative relationships and trust

    Kuldeep Kumar *, Irma Becerra-Fernandez1

    Alvah H. Chapman Graduate School of Business, College of Business, Florida International University, 11200 SW 8th Street,

    Miami, FL 33199, United States

    Available online 27 September 2005

    Abstract

    While relationships and trust are now commonly accepted as central to conducting business both within and across

    organizations, literature provides only minimal guidance as to how relationships and trust in inter and intra-organizations

    are created. Moreover, the role that information technology could play in building trust has largely been ignored.

    This paper describes an Internet-based process for building trust between collaborative commerce partners. Integrating

    concepts from Winograd and Flores concept of speech act theory based bconversations for actionQ with research on bclosedloop cyclesQ of trust and relationships in the disciplines of management and marketing, the paper first develops a frameworkfor understanding trust and trust-building processes. The framework characterizes the process of building of trust as the

    management of commitments inherent in speech acts between requesters (customers) and performers. Furthermore, joint

    sense making during the conversation for action contributes to greater transparency thereby also increasing the levels of

    mutual trust.

    www.elsevier.com/locate/dss0167-9236/$ - see front matter D 2005 Published by Elsevier B.V.

    doi:10.1016/j.ds

    E-mail addresses: [email protected] (K. Kumar),

    [email protected] (I. Becerra-Fernandez).1 Tel.: +1 305 451 2830.2007) 584606Fresh flowers are blooming on the battle-scarred

    landscape where once-bitter rivalries among suppliers,s.2005.05.017

  • customers and competitors once took place.Q (Daft [9],p. 179)

    While collaboration is the sine qua non of any

    organized human endeavor, it has only recently been

    recognized as an important idea underlying business.

    After decades of building upon an adversarial, Por-

    terian2 view of strategy and inter-organizational rela-

    tionships, business and academic press is now

    embracing the religion of trust and relationships

    [31]. Trust is being trusted again.

    Implicit in the Porterian view of inter-organiza-

    tional relationships is a view of the transacting

    partners as opportunistic, and consequently strategy

    K. Kumar, I. Becerra-Fernandez / Decisionas a means for controlling the potential opportu-

    nism of the trading partners. This view of the

    other party as an adversarial opportunist derives

    from and underlies economic theories such as

    transaction cost economics [42], agency theory

    [4], and resource dependency theory [30]. In trans-

    action cost theory and agency theory, transaction

    costs and agency costs are costs deemed to be

    incurred for protecting against and controlling the

    potential opportunistic behavior of the other party

    in the business transaction or, in case of an agency

    relationship, of the firms agents. In resource

    dependency theory, the focus of strategy is on

    minimizing dependency on others while maximiz-

    ing others dependency on us. Hence, the approach

    to managing transaction costs is either to supervise

    closely the other party or agent, or alternately, use

    contracts, incentives, punishments, or technology to

    2 Michael Porters seminal bfive forces modelQ has been thestaple opening for many strategy courses in graduate and under-

    graduate business programs. With his prescription of gaining

    power over suppliers, locking in customers, and creating barriers

    to new entrants, Porter creates a paranoid view of the business

    relationships and business environment. The underlying assump-

    tion is that everyone, our suppliers, our customers, and for sure our

    competitors are out to get us, and we better get them before they

    get us. However, there are internal contradictions in Porters pre-

    scription. This paranoid view of strategy is at odds with his idea of

    an extended value chain and a value system across multiple

    organizations. A value chain requires cooperation and collaboration

    between different units within the organization, a value system

    between different organizations in the supply chain. Porter, like

    many other strategists, takes a control view of the environment.

    Strategy depends on the ability to control what is external to theorganization, that is, its suppliers, customers, competitors and new

    entrants, and new substitute products.monitor and ensure their compliance with the con-

    tract. The former ultimately culminates into the

    total incorporation of the other party or the agent

    into the organizational hierarchy thereby achieving

    full control over them; in the case of the latter,

    instruments of control are designed and implemen-

    ted to align the agents behavior with the firms

    interests. In either case, the key strategic thrust is

    one of controlling the other partys potential

    opportunism:

    bWe agree that control is an overarching issue forbusiness organizations. According to Yates [44], most

    technologies and organizational forms have had as

    their main objective the creation of more advanced

    control instrumentsinstruments that enable us to

    enhance and extend our control over processes in

    society and nature. Correspondingly, most of the

    management literature continues to provide models

    and tools to enhance and support control over busi-

    ness processesproduction, distribution, marketing,

    sales, and so on.Q (Ciborra [6], p. 3)

    In information systems, the Porterian view of

    the inter-organizational relationships manifests itself

    as the much-heralded strategic role of IT for com-

    petitive advantage, and the characterization of IS as

    a competitive weapon, as presented, for example,

    in the work of Applegate, Mc Farlan, and Mc

    Kenny [1]. Likewise, the logical outcome of the

    transaction cost perspective in information systems

    is the proposition that information and communi-

    cation technologies are strategic mechanisms for

    controlling transaction risks and transaction costs

    of necessary, but inherently risky, intra- and inter-

    organizational interdependencies and relationships

    [7,27].

    However, we are beginning to realize that it is

    not always possible to impose complete control over

    our organizations or our business partners [6]. This

    is especially the case in those situations where the

    power between the two parties is evenly matched.

    Even in the case of unequal power, it is not always

    possible to observe, monitor, and control the variety

    of ways in which a subordinate in a hierarchy or a

    trading partner in a business relationship can stray

    from the desired behavior. The notion of incomplete

    Support Systems 43 (2007) 584606 585contracting and consequent need for relational con-

    tracts suggests that not all possible contingencies in

  • cisiona relationship can be anticipated, explicitly con-

    tracted for, and monitored. Moreover, severe pro-

    blems of work specification and work monitoring

    and acceptance can arise when the object of trade or

    collaboration is abstract (such as a business plan or

    R&D effort), uncertain, complex, and in general,

    less specifiable. Furthermore, in the case of interna-

    tional trade and global e-business, two additional

    complications arise [22]. First, global distances

    make it problematic to supervise or to observe

    and monitor compliance at a distance. Second, inter-

    national differences in culture, legal systems and

    jurisdictions, and governance can make the interpre-

    tation and enforcement of contracts difficult.

    Ciborra, in continuing with his discussion of control

    observes:

    bBut we submit that control is difficult to achieve.Nature, society, and the economy have always been

    unpredictable and uncontrollable. Although technol-

    ogy allows us to sharpen our governance capabilities,

    we seem to end up deploying technology to create a

    world that resists control. That is what globalization

    is all about: not just extended transactions or higher

    cross-border investments. We are experiencing gov-

    ernance in the age of globalization is more limited

    than ever.Q (Ciborra, [6], p. 3)

    Consequently, organizations have no recourse but

    to search for alternatives to control. Economists

    suggest control by markets as an option. Extending

    transaction cost economics, Malone et al. [27]

    predict that with information technologys ability

    to reduce transaction costs, increasingly firms

    would be operating in a market. On the other

    hand, large organizations such as automobile com-

    panies and electronic companies are moving

    towards developing intense, long-term relationships

    with a few preferred suppliers. Concepts such as

    supplier and customer relationship management are

    gaining acceptance in business. Keen [21] states

    that relationships and trust, instead of technology,

    drives the growth of e-commerce. Handy [17]

    observed that virtual organizations tend to rely

    more on trust, than their traditional counterparts

    operating in markets.

    This brings us to trust and relationships as an

    K. Kumar, I. Becerra-Fernandez / De586alternative to control. Bradach and Eccles [3] con-

    ceptualize trust as a form of control, together withtraditional forms of control, markets (price mechan-

    isms) and authority (hierarchy). However, the notion

    of trust implies that we are ceding a measure of

    control to the trusted party, thereby actually substi-

    tuting trust for control. Therefore, rather than con-

    sidering trust as a form of control, we prefer to

    characterize trust as a substitute for and complement

    to power and control. Business press and academic

    researchers have discovered that intra and inter-orga-

    nizational transactions do not always exist in an

    atmosphere of distrust and suspicion, thus requiring

    reliance on control [10,18]. There are a number of

    real-world situations, both in the new and the old

    worlds, where business is conducted based upon

    relationship and trust between the trading parties

    [23]. This focus on relationships is a response to

    the idea of opportunism that pervades transaction

    cost economics and agency theory. Even as organi-

    zational theorists have been developing the concept

    of trust and relationships for nearly half a decade,

    the awareness of the role of trust and its relationship

    to technology in information systems literature is

    only recent, dating only back to the work by

    [19,20,23].

    But all this talk about the importance of trust

    does little to help us understand how relationships

    and trust come into being. While the above lite-

    rature examines the meaning of trust, and recog-

    nizes its importance in business relationships, with

    few exceptions, it does not address the question as

    to how trust can be created and sustained [38].

    Decades of dwelling on strategies and tactics for

    guarding against possible opportunism and betrayal

    from our trading partners have left us with very

    little guidance as to how can we create and build

    the trust that now seems so vital to conducting

    business. We either naively assume that an at-

    mosphere of trust already exists, in which case

    business transactions can be initiated, or a lack of

    such pre-existing relationships can leave us at a

    loss for initiating potentially mutually beneficial

    business endeavors. Lacking a pre-existing set

    of relationships, trading partners have no guide-

    lines for initiating business transactions or building

    a business relationship. Despite all the attention

    to trust by leading observers in business and aca-

    Support Systems 43 (2007) 584606deme, very few have actually examined the

    mechanisms and processes by which relationships

  • and trust are created. Finally, the role that tech-

    nology plays in building relationships has been

    largely ignored.

    K. Kumar, I. Becerra-Fernandez / DecisionThe objective of this paper is to develop and

    describe an information technology-based process

    for building relationships and trust between business

    partners. For the purpose of this paper, business

    partners include customers, suppliers, co-operating

    companies, complementors, and competitors, both

    within and across organizations. Such a trust-build-

    ing process will be relevant to both practice and

    theory building. From a practice perspective, the

    guidance provided by this theory-based process

    will provide guidelines and tools for building trust

    in business transactions and relationships.

    From the perspective of research, by examining the

    process of building trust, it will lead to a deeper

    understanding of what constitutes trust. Herbert

    Simon [36] once criticized Barnard for being too

    preoccupied with strategic factors and theories, and

    failing to provide a general treatment of the design

    process.3 Following Simons exhortations, under-

    standing how organizations create new products,

    new methods, new strategies, and new organization

    forms is important. A more fundamental need is to

    understand how organizations create trust that under-

    lies the very basis of sustained relationships within

    and across organizations. In this paper, we embark on

    this challenging task.

    The remaining paper is structured as follows.

    The next Section, 2, provides an overview of the

    current literature dealing with building relationships

    and trust. Section 3 outlines our theoretical frame-

    work for the process of building relationships and

    trust. The next section, Section 4, specifies the

    requirements for information technology support for

    the process of building relationships. Section 5

    describes ActionWorksR Metro, a tool designed formanaging commitments and trust. It is followed by a

    bproof-of-conceptQ case study describing the use ofthis tool in a business setting. The last section con-

    cludes with a summary and potential issues in the use

    of such a tool.

    3 This argument for the need for developing a method for trustcreation emulates the argument by Nonaka and Takeuchi (19xx, p.

    50) for a need for the theory of knowledge creation.2. Previous literature on building trust

    A review of the literature on trust suggests that trust

    and social and working relationships are built over

    time [15]. Dyer and Chu [14] observe that trust

    requires time to develop, and the longer the duration

    and history of interaction, the higher the level of trust.

    Zucker [45] identifies trust as derived from concrete

    experiences resulting from personal relationships that

    are tied to the past (process-based), tied to a person

    (characteristic-based), or tied to societal structures

    (institutional-based). Dwyer et al. [13] distinguish

    five phases in relationship development among trad-

    ing partners: awareness of the trading partner, ex-

    ploration of the benefits and risks of trading,

    expansion to increase personal benefits, commitment

    to continuity in the relationship between partners, and

    durability over time to develop trust. Similarly,

    Gabarro [15] recognizes four stages in relationship

    development: orientation to form an impression,

    exploration of expectations, testing the relationship

    trust, and stabilization of interpersonal contract.

    Lewicki and Bunker [25] point out that trust is easier

    to destroy than to build and that b[while] cooperativeprocesses facilitate trust, some of the types of trust we

    propose can exist in competitive relationships as wellQ(p. 134).

    Many authors liken the process of building trust to a

    closed loop cycle [1315,28,33,45]. They suggest that

    each successful completion of the cycle contributes to

    enhanced levels of trust between the partners. The

    cycle of trust creation is often considered as a loop

    of repeated interactions, in particular, between the

    customers encounter with their suppliers salesperson,

    where trust is built through the repeated delivering by

    the salespeople on their promises. Moreover, the fre-

    quency of business contact positively influences trust,

    which confirms the positive feedback characteristics in

    the process of building trust [14,15].

    Ring and van de Ven [33] describe interorga-

    nizational relationships as starting with small, in-

    formal deals that initially require little reliance on

    trust because they involve little risk. As such trans-

    actions are repeated over time, meeting specified

    norms of equity and efficiency, commitment among

    parties increase, eventually becoming a long-term

    Support Systems 43 (2007) 584606 587bweb of interdependent commitmentsQ (p. 100). Infact, according to Ring and van de Ven: bIncreases in

  • trust between parties, which are produced through an

    accumulation of prior interactions that were judged

    by the parties as being efficient and equitable,

    increase the likelihood that parties may be willing

    describes the positive recursiveness of trust, where the

    cycle of trust begins with ba policy of small stepsQ

    K. Kumar, I. Becerra-Fernandez / Decision588[[38], p. 39] and trust is the result of a positive experi-

    ence with persons via dfaceworkT commitments inexchange relations, with bfrequent, repeated, andmultifaceted contacts among organizations and an

    open exchange of informationQ (p. 48) resulting inincreased trust.

    While the above body of literature does point

    to a sustainable cyclical strategy for developing

    and maintaining trust, it is silent on the precise tactics

    and steps as to how this cycle can be created and

    managed. Moreover, it does not address the role of

    information technology in sustaining this cycle. When

    organizations and persons transact business at a dis-

    tance, information and communication technologies

    are key to mediating communication between the

    transacting parties. Thus, it is important that the role

    of these technologies is addressed in designing trust

    creating and sustaining processes.

    3. On framing the building of trust and

    relationships

    Before we discuss the process of bbuilding trust,Qwe need to be clear about the meaning of trust itself.

    The literature conceptualizes trust in two ways. First,

    and most commonly, trust is thought of as a bnounQ ormore precisely, a bstate-variable.Q As a state-variable4,

    4 Here we use the term state-variable in the sense that it is a

    variable that describes the state or level of the state attribute. Thus itto make more significant and risky investments in

    future transactionsQ [[33], p. 100].Although the phases involved in the trust-building

    loop may vary somewhat among authors, the com-

    monly accepted repetitive and positive feedback char-

    acteristics of the process of building trust indicate that

    successful cooperative relationships as described by

    this loop result in increased trust [32,33]. Sydow alsois analogous to the idea of a level in the terminology of Foresters

    System Dynamics.it is an attribute or a property with variable values, that

    is attached either to the trusted party, or to the relation-

    ship between the trusting-party and the trusted party.

    The former deals with the btrustworthinessQ of thetrusted party. The latter results in statements such as

    bthe customer has trust in the supplier,Q or bI havetrust in you.Q It is reflected in statements such asbmy trust has to be earnedQ or btrust was lost.Q Trustis thus conceptualized as an object like a building or

    an automobile that the trusting-party either has or

    owns, the trusted party earns, or either can lose.

    While in these illustrations, trust is clearly intended

    as a noun, as we will show later, conceptualizing

    trust as a state-variable is a more precise and useful

    conceptualization of the concept.

    The second conceptualization of trust is that of

    a transitive bverbQ as in bIn God we trustQ or inbFirestone does not trust Ford anymore.Q In thesecases, trust is taken to be an actiona process-based

    view of trust. Trusting is something you do as opposed

    to something you have. As a transitive verb, it has both

    a subject (the trusting-party) and an object (the trusted

    party) of the action. Trust is thus an interaction

    between two parties. Following the ontology of

    entity-relationship modeling, the verb (the act of trust-

    ing) establishes the relationship between the subject

    and the object roles. This brings in an added complica-

    tion. The roles of the subject and object are not fixed.

    They are reciprocal. The designated subject A, in

    acting, interacts with the designated object B. As

    interactions are often reciprocal (not reflexive), while

    party A acts on party B, B is concurrently acting on A.

    Thus, usually an interaction results in party A trusting

    (or not trusting) party B and at the same time B trusting

    (or not trusting A). Furthermore, the strength of As

    trust in B need not be the same as the strength of Bs

    trust in A.

    Our assertion is that the verb and the state-variable

    views of trust are complementary. The act or the

    process of trusting during an interaction, together

    with the outcome of the interaction, leads to an

    increase or decrease in the level of trust. Now trust

    the noun is no longer static like a building. Its value

    changes as the result of the interaction. It is, therefore,

    a dynamically changing bstate-variableQ in a state-transition where trusting during the interaction creates

    Support Systems 43 (2007) 584606the transition between different levels or states of

    trust. Sydow [[38], p. 3435], in the context of

  • inter-organizational trust, uses Giddens concept of

    structuration to develop this duality5:

    bFrom a structuration perspective one would preferthe notion of trust constitution because, in contrast to

    other concepts, it emphasizes both the possibility of

    intentional creation and the emergent development of

    trust and, in particular, the subtle interplay of these

    two dimensions of the constitution process. Trust,

    even if attributed to certain personal or organizational

    characteristics, is mainly produced and reproduced

    via action, in the case of inter-organizational net-

    works, via management interaction in particular.Q

    K. Kumar, I. Becerra-Fernandez / Decision(Sydow [38])

    As we are interested in the problem of building

    relationships and trust, we take a state-transition or a

    process perspective on trust. This view of trust

    builds upon Ring and van de Vens [32,33] perspec-

    tives on building relationships, Solomon and Flores

    [37] concept of bBuilding Trust,Q and Zuckers [45]concept of bprocess-based trust.Q It is also compati-ble with Luhmans [26] functionalist view where

    trust is taken as a mechanism that reduces the inter-

    nal complexity of interactions.6 Zuckers concept of

    process-based trust which is tied to past and

    expected exchanges entails the incremental process

    of building trust through gradual accumulation of

    knowledge [24]. This bexperiential knowledgeQ devel-ops from a concrete experience of social or economic

    exchange and is brought as an expectation to future

    transactions [45].

    Ring and van de Ven [33] suggest that the process

    of building inter-organizational relationships, as illu-

    6 For Luhman, trust is a mechanism by which actors reduce the

    internal complexity of their system of interaction through the adop-

    tion of specific expectations about the future behavior of the other.

    Luhmanns functional approach bypasses the question of the grounds

    of trust. By emphasizing the role of expectations, he implies the

    importance of cognitive structures in as far as shared meanings

    reduce complexity in a social interaction. Trust(ing) absorbs com-

    5 Sydow (p. 3739) goes further in adapting Giddens structura-

    tion theory to establish a duality and recursiveness between action

    and structure as both sources and outcomes of trust. While this in-

    depth understanding of the relationship between trust, action, and

    structure is illuminating, it is not needed to understand our argument

    and is not developed here.plexity in so far as someone who trusts acts as if the trustees actions

    are, at least to some degree predictable (Lane 2000, p. 1213).strated in Fig. 1, is one of gradual development and

    evolution, and consists of a repetitive sequence of

    cycles made of negotiation, commitment, and execu-

    tion stages. They further suggest that relationships are

    products of numerous such interactions and thorough

    these interactions trust emerges.

    On the other hand, Solomon and Flores [37] main-

    tain that trusting is a means to creating, maintaining,

    deepening, and restoring relationships. Thus, between

    Ring and van de Ven [33] on one hand, and Solomon

    and Flores [37] on the other, we have a reciprocal

    interaction between relationship and trust. Conse-

    quently, we interpret the interaction cycle proposed

    by Ring and van de Ven as a means for building both

    relationships and trust.

    However, Ring and van de Vens interaction cycle,

    consisting of negotiations, commitment, and execu-

    tion, is not complete. A transaction may only be

    considered complete when the parties to the transac-

    tion either accept the outcomes as satisfactory, rework

    until satisfaction is achieved, or withdraw from the

    transaction [[43], p. 6466]. Declaring acceptance by

    the requestor or withdrawal by either party is the

    condition of completion. It is this completed transac-

    tion that leads to Zuckers experiential knowledge,

    that, in turn, leads to increased (or decreased) levels

    of trust.

    In addition to the experiential nature of building

    trust (trust as a product of accumulation of knowl-

    edge about outcomes of past transactions), the very

    process of interacting, that is, the process of negotia-

    tion, conversation, commitment, performance, and

    acceptance, by itself, also creates trust. Both Ring

    and van de Ven [33] and Solomon and Flores [37]

    come to this conclusion from somewhat different

    directions. Ring and van de Ven suggest that in the

    negotiations stage, the parties develop joint (not

    individual) expectations about their motivations, pos-

    sible investments, and perceived uncertainties in the

    business deal [33]. Following the arguments of Com-

    mons [8] and Turner [39], they suggest that the

    development of inter-organizational relationships is

    grounded in the predispositions of the interacting

    parties to engage in mutual sense making and bond-

    ing. These sense-making and bonding processes per-

    mit parties, with initially different views of their

    Support Systems 43 (2007) 584606 589potential purposes and expectations of a relationship,

    to achieve congruency in their relationship. Accord-

  • CUTIO

    ments th

    SSME

    ased on

    fficienc

    equity

    lopm

    cisionNEGOTIATIONS

    of joint expectationsrisk and trust through

    formal bargaining

    informal sense making

    EXE

    of commit

    ASSE

    b

    e

    Process Framework of the Deve

    K. Kumar, I. Becerra-Fernandez / De590ing to them, sense making is an enactment process

    [40] in which organizational participants come to

    appreciate the potential for transacting with others

    by shaping and clarifying the identity of their own

    organization in the context of the interaction.

    bInteraction is based upon the need among transact-ing parties to feel that they share a congruent under-

    standing of an inflexible world. Communications

    among parties produce this shared interpretation,

    and it often emerges gradually and incrementally.

    Congruency is a cumulative product of numerous

    interactions; through these interactions emerge trust

    in the good will of others and an understanding of

    constraints on the relationship.Q (Ring and van deVen [33])

    Solomon and Flores [37] come to the same under-

    standing from the perspective of authenticity. Their

    notion of bAuthentic TrustQ is borrowed from theexistentialist tradition, particularly from Kierkegaard

    and Heidegger. In this tradition, authenticity involves

    role interactio

    personal interac

    Fig. 1. Ring and Van de Ven diagCOMMITMENTS

    for future action through

    formal legal contract

    psychological contract

    NS

    rough

    NTS

    :

    y

    ent of Cooperative IORs

    Support Systems 43 (2007) 584606a keen awareness, both of ones own identity and of

    ones relationships with others, and a deep awareness

    that self-identity is fluid and uncertain and that our

    identities change with our circumstances and our

    commitments to others (Solomon and Flores [37],

    p. 91). Thus, interaction and trusting become a pro-

    cess of mutual self-discovery for the transacting

    parties. This discovery comes about by articulating

    and clarifying the object of the interaction in a dialog

    or conversation about it:

    bPeople do not develop trust by forming affectiveattitudes or beliefs about another person. They

    develop trust through interaction and conversation in

    relationships with each other. Authentic trust by its

    nature is articulated trust, trust that is bspelled out.Q Assuch it becomes an issue. This has double signifi-

    cance. For both parties, becoming aware of their

    obligations and responsibilities intensifies their sense

    of mutual identity and the significance of the relation-

    ship for each of them. In business contexts trust is

    always an issue. Whether it is in the quality of product

    ns

    tions

    ram process network [33].

  • 3A: Declare

    A:

    t B

    on (fr

    ision Support Systems 43 (2007) 584606 591or the promptness of the delivery,whether it is

    respect for the importance of confidential information

    or the need for joint strategy, the issue is trust, which

    will be articulated and negotiated, whether or not in

    the form of an explicit verbal agreement or a written

    1 2

    8

    2

    A: Request B: Promise

    A: Withdraw

    B: Reject

    A: Reject

    B: Withdraw

    A: Counter

    B: Counter

    A: Accep

    Fig. 2. The basic conversation for acti

    K. Kumar, I. Becerra-Fernandez / Deccontract.Q (Solomon and Flores [37], p. 9697)

    At this point, we also need to recognize that this

    process of mutual discovery and sense making is

    limited, not only to the negotiation phase of the

    transaction. The dialog and exchanges dealing with

    the clarification of requests, negotiations, commit-

    ment, and even the dialog during the acceptance

    stage gradually increase the levels of mutual and

    joint understanding through out the evolution of the

    interaction. Trust and relationships evolve not only

    through repeated cycles of interaction, but also

    throughout the various stages of the same interaction

    cycle.

    Thus, from the perspective of both sense making

    and authenticity, the transacting parties participate in

    interactions through which they come to a mutual

    understanding of their own and the others identities

    and expectations. This interaction is manifested

    through a dialog or conversation that takes place

    between the transacting parties. Request for service,

    promises to perform the service as requested, theassertion that the performance7 is concluded and the

    promise fulfilled, and the acceptance of the outcome

    as satisfactory, are performed as linguistic acts entered

    into by the requestor and the performer.

    Consequently, an interaction or an exchange can be

    4 5

    7

    9

    B: Assert A: Declare

    A: Withdraw

    Withdraw

    : Renege

    om Winograd and Flores [43] p. 65).characterized as a conversation in which parties to the

    exchange go through cycles of request; negotiation,

    promises and commitment; statement of performance;

    and acceptance. Winograd and Flores state that there

    is a recurrent structure to these conversations acts that

    can be formalized into a network of speech acts

    shown in Fig. 2.

    In Fig. 2, each circle represents a possible state of

    conversation and the arrows between them represent

    speech acts. The diagram outlines a network of speech

    acts that constitute conversations for actionconver-

    sations in which in which an interplay of requests and

    commissives are directed towards explicit cooperative

    action (see Appendix A for an explanation of the

    concept of speech acts).

    The fundamental importance of the illocutionary or

    speech act is that it embodies the specification of mean-

    ing in terms of patterns of commitment entered into by

    7 While performance, such as the delivery of a product or service,

    is physical act, the assertion by the performer that it has been

    completed is a language or a speech act.

  • useful, they leave out the essential dimension of col-

    laborative work. Thus, they do not fully exploit the

    cisionthe speaker and the hearer by virtue of their taking part

    in the conversation. Habermas observes the centrality

    of commitments in speech or illocutionary acts:

    bThe essential presupposition for the success of anillocutionary act consists in the speakers entering into

    a specific engagement so that the hearer can rely on

    him. An utterance can count as a promise, assertion,

    request, question, or avowal, if and only if the speaker

    makes an offer that he is ready to make good insofar

    as it is accepted by the hearer. The speaker must

    engage himself,Q [[16], p. 61].

    The engagement inherent in the speech act based

    conversation leads to commitment, and repeated meet-

    ing of commitments leads to experiential or process-

    based trust:

    bTrust is built step by step, commitment by commit-ment, on every level. Trust must be built one step

    (sometimes it seems like a giant step indeed) at a time,

    by the way of interpersonal confrontations and mutual

    engagements, by the way of commitments and pro-

    mises, offers and requests.Q (Solomon and Flores,[37], p. 49)

    The step-by-step process of building commitment,

    relationships, and trust can be visualized as a lin-

    guistic dance between the interacting parties.

    Requests and offers and counter-offers, assertions

    of completion, and declarations of satisfaction,

    require the use of language to perform the speech

    acts:

    bBut the essence of building trust is making commit-ments, and wordless commitments are rare. Indeed

    making and honoring commitments involve precisely

    the same combination of words and action that build

    trust.Q [[37], p. 36]

    Thus, a series of formalized patterns of speech acts,

    i.e., bconversations for action,Q embody the trust andrelationship-building process. Trust and relationships

    are created by both the ongoing process of creating

    commitments resulting in common mutual under-

    standing of expectations, capabilities, and constraints

    of each other and the experience-based knowledge of

    meeting (or not meeting) these commitments.

    As a final point, it should also be recognized that

    K. Kumar, I. Becerra-Fernandez / De592during a conversation a variety of bbreakdownsQ couldoccur. For example, the performer may not preciselypotential of computers.

    Winograd and Flores argument can also be

    extended to the current generation of computer-

    based systems such as enterprise resource planning

    (ERP) systems, supply chain management systems

    (SCM), and even customer relationship management

    (CRM) systems. Although business process models

    embedded in ERP and SCM systems are effective for

    observing the movement of information and material

    across the organization:understand the request. Or, after the performer declares

    completion, the originator of the request may not agree

    that the outcome is acceptable. In these cases, further

    conversations between the requestor and the performer

    can take place to resolve the breakdown. The speech

    acts used to resolve these breakdowns at any point in

    the cycle in the transaction are also part of the process

    of interaction and sense making, and therefore can

    contribute to the process of mutual common under-

    standing and, if resolved satisfactorily, ultimately a

    strengthened relationship and increased trust.

    4. Requirements for collaborative technology

    support for building trust and relationships:

    tools for conversation

    Based upon the concepts from Winograd and

    Flores [43], Solomon and Flores [37], and Ring and

    van de Ven [33], in the previous section, we estab-

    lished bconversations for actionQ as central to trustand relationship building. In this section, we discuss

    the role of computers and information technology in

    supporting these conversations.

    In their seminal work, Understanding Computers

    and Cognition, Winograd and Flores [[43], p. 157]

    argue that traditional applications of computers such

    as decision support systems are bnot the most pro-mising domain in which to build tools for managing.QThey observe that many systems designed by compu-

    ter professionals are intended to facilitate the activity

    of an individual working alone. Although such tools

    (including word processors, filings systems etc.) are

    Support Systems 43 (2007) 584606bthese models are blind to the human process in whichpeople request work and agree on what will be done,

  • plex as we can no longer rely on previously developed

    mutual understanding and shorthand for communica-

    tion. In the case of global trade and e-business, the

    lack of a common cultural base, global gaps8, and

    ision Support Systems 43 (2007) 584606 593who will do it, and when it will be done; they provide

    no mechanism for ensuring that any customer is

    satisfiedQ (Denning and Medina-Mora [12], p. 45).

    While ERP and SCM systems operate at an orga-

    nizational or inter-organizational level, they merely

    connect individual bdesktopsQ to the system, not peo-ple to people. In such systems, individual persons still

    work alone, albeit now embedded in pre-defined pro-

    cess-structures and interactions with the system.

    On the other hand, collaboration and therefore the

    conversational dimension permeate every realm of

    coordinated organizational activity, whether it is com-

    puter programming, medical care, or selling flowers.

    The details differ from setting to setting, but there is a

    common theoretical basis and a common regular

    structure of these conversations (Winograd and Flores

    [43], p. 158).

    Winograd and Flores suggest that a more promis-

    ing avenue for utilizing the potential of computers

    would be the use of computer-based tools for sup-

    porting conversations for action. Their basic argu-

    ment is that organizations exist as networks of

    speech acts, primarily directives and commissives.

    Directives include orders, requests, counter-requests,

    consultations, and offers. Commissives include pro-

    mises, acceptances, and rejections. Furthermore,

    breakdowns in the process of interaction inevitably

    occur. In coping with breakdowns, further networks

    of directives and commissives are generated. To

    develop the conversations consisting of patterns of

    these directives and commissives people issue utter-

    ances by speaking or writing. They participate in the

    creation and maintenance of a process of commu-

    nication and joint sense making. At the core of this

    process is the performance of linguistic (or speech)

    acts that establish different types of commitments.

    Based on Solomon and Floress and Ring and van de

    Vens argument, the process of joint sense making

    and the experience of satisfied commitments build

    trust and relationships.

    The networks of commitments and the conversa-

    tions in which people participate become wider in

    scope as we venture into inter-organizational value

    production systems. Moreover, if the inter-organiza-

    tional interaction is among parties who were strangers

    K. Kumar, I. Becerra-Fernandez / Decbefore the initiation of the interaction, the conversa-

    tion for joint sense making becomes even more com-poly-contextuality make these conversations even

    more complex [23].

    The complexity of the conversation also increases

    as often a part of the work needed to perform on the

    commitment is delegated or outsourced to another

    person, organizational sub-unit, or another organiza-

    tion. Thus, a commitment to perform some work may

    generate further commitments, which in turn may

    generate further sub-commitments, resulting in a net-

    work of cascading commitments. The satisfaction of

    a particular commitment then is contingent on the

    satisfactory performance of all its cascading sub-com-

    mitments. Moreover, any breakdowns in sub-commit-

    ments travel upstream the cascade thereby increasing

    the complexity of the breakdown monitoring and

    resolution process.

    It may argued that traditional communication

    tools, such as phone-calls, letters, faxes, and e-mails

    can be used to manage these conversations for action.

    However, in todays business interactions and rela-

    tionships, the complexity of such conversations has

    gone beyond the point where it can be managed

    without appropriate tools. In other than the simplest

    interactions, the task of remembering the structure of

    roles and associated responsibilities in a business

    transaction, contexts of interactions, and the cascad-

    ing networks of speech acts and commitments and

    responding to them through appropriate reactions and

    speech acts is beyond the cognitive limits of human

    information processing. Moreover, given the com-

    plexity, participants in the interaction often do not

    (or cannot) specify the roles, interactions, commit-

    ments, and express their satisfaction (or dissatisfac-

    tion) with the completion, with adequate explicitness,

    completeness, and specificity. Winograd and Flores

    maintain that it is in the realm of supporting such

    human interactions where computers can provide the

    most advantage.

    8 Kumar and van Fenema (2001) recognize six such gaps or dimen-

    sions of context: geographical gap, time gap, infrastructure gap,

    culture gap, governance gap, and regulatory gaps. These gaps createdistances in understanding and expectations, distances that need to be

    overcome through conversations aimed at joint sense making.

  • legal and financial conditions required by the transac-

    tion [41]. Collaborative commerce software is based

    cisionIn order to realize this, Winograd and Flores

    exploit the theoretical basis of illocutionary acts and

    the recurrent structure of conversations for action.

    They observe that the rules of conversation are not

    arbitrary conventions like the rules of chess, but

    reflect the basic nature of human language and action:

    bThere are surprisingly few basic conversationalbuilding blocks (such as request/promise, offer/accep-

    tance, and report/acknowledgement) that frequently

    recur in conversations for action. The development

    of a conversation requires selection among a certain

    finite set of possibilities that is defined by the opening

    directive and the subsequent responses. It is like a

    dance, giving some initiative to each partner in a

    specific sequence.Q [[43], p. 159].

    The taxonomy of speech acts and the diagram of

    conversation structure presented in Fig. 2 deal with

    the fundamental ontology of linguistic acts. Winograd

    and Flores assert that this taxonomy and the ontology

    provide a basis for design of tools to operate in the

    linguistic domain of conversations for action [[43], p.

    158159]. These tools can be used in requesting,

    creating, and monitoring commitments. Using the

    recurrent structure of conversation, they can provide

    relevant answers to the questions bWhat can I (or do Ineed to) do next?Q By scanning the network of com-mitments represented in the tool they can answer,

    bWhat is the status of my active and unfulfilledcommitments?Q or bbreakdowns in which sub-com-mitments are likely to affect the performance of

    dependent commitments?QWinograd and Flores are careful in stating that they

    are not suggesting that the computer can bunderstandQspeech acts by analyzing natural language utterances.

    What they are suggesting is that the parties to the

    conversation can be made aware of the structure of the

    conversation and be provided tools for working with it

    explicitly. The conversationalists, through a software

    system on a platform of computers and communica-

    tions network, can share these tools and the tool-based

    representation of the network of speech acts. The

    objective of the computer system would be to make

    the interactions transparentbto provide a ready-to-hand tool that operates in the domain of conversation

    K. Kumar, I. Becerra-Fernandez / De594for actionQ (Winograd and Flores [43], p. 159).According to them, the functional requirements forupon the taxonomy of speech acts ("N41) and exploitsthe conversation structure (Fig. 2) described in Sec-

    tion 3 of the paper.

    In ActionWorksR Metro, the structure of speechacts in the conversation is organized into a closed loop

    structure, called the Business Interaction Model,

    shown in Fig. 3. The closed loop structure is based

    upon the recurrent structure of conversations shown in

    Fig. 2 above. Denning and Medina-Mora [12]

    describe the basic elements of this closed loop:

    bA closed loop. . .that connects two parties. One ofthem promises to satisfy a request of the other. . .Theloop consists of four stages separated by four speechsuch a tool include facilities for speech act origination

    and identification; monitoring completion; keeping

    temporal relationships between a network of cascad-

    ing commitments; and periodic or on request exam-

    ination of the network to show status of various

    commitments. Winograd and Flores specification of

    these requirements is excerpted in Appendix B of this

    paper. By embedding these requirements for the struc-

    ture of the conversation in a computer-based tool, we

    can ensure that the conversations are carried out in a

    manner such that all participants to the interaction

    have a common understanding of the status of the

    interaction, and the status of the interaction can be

    monitored to completion.

    5. ActionWorksR Metro: technology support forconversations for action

    The above requirements for a tool for supporting

    conversations for action are implemented through

    ActionWorksR Metro. ActionWorksR Metro is aweb-based collaborative commerce software devel-

    oped by Action Technologies (www.actiontech.com),

    a firm founded by Terry Winograd and Fernando

    Flores [11]. Collaborative commerce software, also

    called interaction technology, coordinates human

    interaction in the uncertain business of inventing,

    designing, developing, deploying, and supporting

    new products and services, while negotiating the

    Support Systems 43 (2007) 584606acts. . .First the customer make a request of the per-former. . .Second, they negotiate on the conditions

  • Negotiation: The Customer and Performer come to Agreement

    agritionfact

    tt

    DeclDeclComCom

    of the

    ision Support Systems 43 (2007) 584606 595that will satisfy the customer. . .Third, the performerdoes the work and ends by declaring that it is done.

    Fourth, the customer accepts the work and declares

    satisfactionQ (p. 45).

    This closed loop structure in Fig. 3 is similar to

    Ring and van de Vens process framework for the

    development of cooperative inter-organization rela-

    tionships as described in Fig. 1. Ring and van de

    Ven observe that the process of building inter-organi-

    Preparation: Customer specifies work to be performed (or Performer offers to do work)

    Acceptance: The Customer evaluates the work and declares satisfaction

    Customer

    PartiesCondSatis

    RequesReques

    DeclareDeclareSatisfiedSatisfied

    Fig. 3. Basic elements

    K. Kumar, I. Becerra-Fernandez / Deczational relationships is one of gradual development

    and evolution, and consists of a repetitive sequence of

    cycles made of negotiation, commitment, and execu-

    tion stages [32,33]. They suggest that relationships are

    products of numerous such interactions and thorough

    these interactions trust emerges.

    However, there are two minor and one major

    differences between the Ring and van de Ven cycle

    and ActionWorksR Metro closed loop. First, theformer separates the phases of bnegotiationQ andbcommitmentQ whereas the latter combines theminto a single phase. This difference is easily resolved

    when one considers the nature of a speech act as

    conceptualized by Winograd and Flores. A negotia-

    tion consists of a series of utterances or speech acts.

    As discussed in Section 3, a speech act, through

    engagement, embodies commitment (also [16]).

    Thus ActionWorksR Metros negotiation phaseincludes Ring and van de Vens both negotiation

    and commitment phases. The second minor differenceis the inclusion of the bpreparationQ of the customersrequest in ActionWorksR Metro as the first phase ofthe loop. This too is easily accounted for, as Ring and

    van de Ven take this phase to be as part of their

    opening of the negotiation phase.

    The key difference between the two closed loops is

    the inclusion of the bAcceptanceQ phase in theActionWorksR Metro loop. According to Winogradand Flores, a transaction can only be considered com-

    Performance: The Performer fulfills the work and reports completion

    Performer

    ee on s of ion

    PromisePromise

    are are pleteplete

    coordination process.plete when the parties to the transaction either accept

    the outcomes as satisfactory, rework until satisfaction

    is achieved, or withdraw from the transaction.9 It is

    this successful completion that closes the loop and

    leads to strengthening of the relationship and asso-

    ciated increases in trust levels. When each work pro-

    cess loop ends with customer satisfaction, stronger

    relationships and increased trust are fostered among

    transacting parties. In order to carry out their commit-

    ments, the performer may be the customer of others to

    whom she or he delegates or out sources work. In this

    way, performers and customers are engaged in a net-

    work of tasks and commitments necessary to complete

    the original request. On the flip side, breaks in the

    loop blead to negative outcomes, such as distrust and

    9 Perhaps the idea of acceptance is implicit in Ring and van de

    Vens execution phase. However, making it explicit as in the

    ActionWorksR Metro cycle makes it certain that this importantphase of completion is explicitly and consciously accounted for

    through appropriate speech acts.

  • Holly Ross requested on 12/12/2000 02:25 PM Hello Kathryn, We need to create a new custom gift book for our customers and we would like you to bid on it. We need 14,000 books chronicling the history and growth of the Pirelli Tire Company. Wed like each book embossed (on the front cover) with the image of the Pirelli P6 tire with Borrani wire wheel. Each of our North American tire dealers will be getting one of these. Please see the attached photo file and history archive file as well as the specs on the wheel design. Our budget is $140,000 and we need them by March 22, which is the date of our North American dealer meeting. Can you do it under these conditions? Holly

    Due date set to 13/Dec/00 05:00P Start date set to 13/Dec/00 09:00A Effort set to 60 hours

    a

    Kathryn McGovern commented on 12/12/2000 02:48 PM We'd love to get a chance to work with you on this exciting project. We will need to look carefully at the drawings, text and color pictures; with our current backlog and your deadline, we will need to work overtime to get this done. I can say the minimum price would be $215,000 and that's still conditional to our full agreement. Holly Ross followed-up on 12/12/2000 02:49 PM If you can do it for that price, and to our specifications, I have received the extra budget. However, I can't go any higher.

    Due date changed to 22/Mar/01 05:00P Kathryn McGovern agreed on 12/12/2000 02:53 PM Wed be happy to do it at that price and delivery date.

    30 additional hours requested

    b

    Kathryn McGovern reported work done on 3/12/2001 02:57 AM This is done. Please note that we did not emboss the books.

    Due date changed to 22/Mar/01 05:00P 80 hours reported for manufacturing on 12/Dec/00

    c

    Holly Ross did not accept work done on 3/13/2001 07:59 PM The due date of this work item is closed; please consider it in your plans. I'm sorry Kathryn, but this will not work. We need the books embossed as agreed. We did not agree on eliminating the embossing. We consider this to be a key design feature for thenot agree on eliminating the embossing. We consider this to be a key design feature for the book.

    Kathryn McGovern reported work done on 3/17/2001 07:24 AM Sorry for the oversight, I have embossed the books as you specified. We are losing moneyon this contract but it was our fault. Even so, we want you to be satisfied and to continue to do business with us.

    20 hours reported for Re-work on 12/Dec/00HOLLY ROSS ACCEPTED WORK DONE AS-IS ON 3/18/2001 09:04 AM Thanks, Kathryn! I received them today and I am completely satisfied. You can be sure you will be our primary supplier of high-end four-color gift books.

    d

    K. Kumar, I. Becerra-Fernandez / Decision Support Systems 43 (2007) 584606596

  • damaged reputation, both of which can render an

    organization ineffectiveQ [[12], p. 55].In the bBusiness Interaction Model,Q the person

    against which completion is assessed in the accep-

    tance phase.

    ne by

    K. Kumar, I. Becerra-Fernandez / Decision Support Systems 43 (2007) 584606 597making the requests assumes the role of the customer,

    while the person doing the work takes on the role of

    the performer. The model is a bclosed loopQ becausethe customer starts the loop of business interactions

    that becomes closed when the customer declares satis-

    faction. The customer role decides and accepts work

    as being complete; satisfaction with the work done is

    evaluated against explicit conditions. Business inter-

    actions that end with customer satisfaction are

    expected to result in an stronger relationships and

    increased trust.

    The Business Interaction Model describes the rela-

    tions between customers and performers based on the

    following concepts:

    A strong definition of rolesevery agreement

    clearly defines who is the customer and who is

    the performer.

    Maintaining the context of all business interactions

    necessary to achieve fulfillment of the customer

    requestthe context defines the current status of

    the interaction, what actions have been performed,

    what actions come next, and who is required to do

    them. The context is based upon the precedence

    relationships identified through the recurrent struc-

    ture of conversations outlined in Fig. 2. The tool

    helps its users in structuring their conversation.

    Using the precedence relationships in Fig. 2, at

    the end of each utterance or speech act, the system

    helps the user select the speech acts that can logi-

    cally follow it to completion. For any commitment

    in the network, it also keeps track of the depen-

    dencies between it, its delegated or outsourced

    commitments, and the commitments that depend

    on it. Thus, the software tool maintains the com-

    plete context of the business interaction.

    Conditions of satisfaction are explicitly specified so

    that interactions are focused on completely satisfy-

    ing the customerthese conditions negotiated in

    the negotiation phase provide the benchmark

    Fig. 4. (a) Preparation: customer (Holly Ross) proposes work to be doperformer come to an agreement about the work to be performed. (c) Per

    Acceptance: customer evaluates the work and declares satisfaction.Here, we use a simple example to illustrate the use

    of the business interaction model. Fig. 4a through d

    illustrate how the technology assists the interaction

    between the customer (Holly Ross), and the perfor-

    mer (Kathryn McGovern) [41]. To keep the illustra-

    tion simple, only a single loop with no subsidiary

    loops is presented. In Fig. 4a, the preparation phase,

    the customer proposes the work to the performer, in

    this case a new custom gift book. In Fig. 4b, the

    negotiation phase, the customer and performer

    negotiate and agree on the work to be performed,

    and the conditions of satisfaction are explicitly spe-

    cified including cost and delivery time. The delivery

    date could be explicitly specified as a result of the

    negotiation, or the process could have a predeter-

    mined cycle time, say for example 14 days, and the

    due date would be specified by the system as 14 days

    after the start date. The speech acts, especially the

    assertives, directives, and the commissives have com-

    mitment inherent in them. In Fig. 4c, the performance

    phase, the performer performs the work and reports

    completion. In Fig. 4d, the acceptance phase, the

    customer evaluates the work done against the specific

    conditions of satisfaction expressed in the negotiation

    phase. This phase closes the business interaction

    loop, and by ensuring the customers conditions of

    satisfaction are met, facilitates the building of trust

    over time. Moreover, the process of interaction dur-

    ing the four phases increases the mutual awareness

    and understanding, building relationships [33] and

    authentic trust [37].

    People learn to trust others by observing the

    consonance of their actions with their speech acts;

    for example, promising to do something and fulfill-

    ing the promise earns trust between transacting par-

    ties [29]. Action Technologies ActionWorksR Metroenables trust building by repeatedly asking the cus-

    tomer, bare you satisfied that we (the performer)have met the conditions we agreed on at the time

    we negotiated this agreement?Q This explicit and

    the performer (Kathryn Mc Govern). (b) Negotiation: customer andformance: performer performs the work and reports completion. (d)

  • recurrent questioning, for every interaction, exposes

    shortcomings in the relationship, creates opportu-

    ness and reducing expectations ambiguity. In that pro-

    cess, it builds trust among collaborators over time.

    K. Kumar, I. Becerra-Fernandez / Decision598Organizations lacking such a technology often find

    themselves relying on traditional communication tech-

    nologies, such as e-mail, to accomplish this task. Tra-

    ditional communication technologies have no notion

    of accountability, and no mechanism to track against a

    specific task or project [11]. Moreover, they leave the

    important task of structuring and ensuring the comple-

    teness of the conversational elements to people subject

    to the cognitive limits of human information proces-

    sing. Therefore, simple communication technologies

    are not adequate to the tasks of structuring interac-

    tions and tracking commitments, processes, and pro-

    jects. Furthermore, ActionWorksR Metro technologycan connect commitment cycles together in a cascad-

    ing effect (also called dependencies) so that changes

    in once cycle, for example delivery time, will auto-

    matically propagate to interconnected dependencies

    [11,12].

    6. ActionWorksR in action: proof of concept atCCPS

    In this section, we illustrate the use of the

    ActionWorksR Metro closed loop in managing thecommitment cycles in the new product development

    process for International Truck and Engine Corpora-

    tion.10 International Truck and Engine Corporation

    manufactures medium-duty to heavy-duty trucks and

    school buses, has around a 40% market share and is in

    10 http://www.internationaldelivers.com/, a subsidiary of Navistarnities to repair dissatisfaction and produces relation-

    ships that are built on trust.

    The value of the ActionWorksR Metro technologyis in first assisting collaborators to structure the con-

    versations for explicitness, completeness, and specifi-

    city, and then keep track of what they have committed

    to be accountable for, what they have delivered, and if

    the customer is satisfied. It helps further by assisting

    the interacting parties in clarifying poorly defined

    commitments thereby both increasing mutual aware-International Corporation (NYSE symbol NAV) a global leader in

    the manufacture and sales of trucks, buses, and diesel engines.the top three among its competitors. However, in the

    fourth quarter of 2002 Navistar faced a net loss for the

    year, as a result of decreased demand and increased

    competition. The industry had a massive global over-

    capacity and was facing a possible consolidation

    among the top ten players, resulting in each player

    taking aggressive moves to cut costs primarily

    through productivity improvements.

    To maintain its dominant market position, Interna-

    tional needed to launch a new product development

    process to speed the introduction of new truck models,

    reduce rework, and cut costs significantly. Tradition-

    ally, Internationals assembly line was marked by a

    needless proliferation of the possible number of

    powertrain combinations, which pushed the individual

    vendor components quantity down and consequent

    price up. The new business process redesign at Inter-

    national focused on reducing this complexity by limit-

    ing the number of combinations available, buying

    components in bigger volumes, and focusing colla-

    boration efforts with suppliers on improved perfor-

    mance issues. As summarized by Navistars CEO

    John Horne:

    bOur current medium truck has more than 800 com-binations of engines and transmissions. Its nuts. The

    new truck has 34 combinations, and every one of

    them works better than before because our engineer-

    ing people had the time to really develop them so they

    feel good to the driverQ [5]

    In order to accomplish this goal, business process

    management became a strategic priority at Interna-

    tional. In particular, the product change management

    (PCM) process, associated with new product devel-

    opment, had to be dramatically improved in terms of

    quality, speed, and reliability. The PCM process

    needed a better way to collaborate, not only intern-

    ally, but with customers and suppliers as well. They

    needed a solution that would reduce cycle times,

    ensure quality (i.e., applications that met specifica-

    tions and deadlines), improve productivity, and

    enhance coordination and relationships with custo-

    mers and suppliers.

    The PCM process involves about 1000 employees

    in production facilities located across the US, Mexico,

    and Canada, as well as external suppliers. Since Inter-

    Support Systems 43 (2007) 584606national had dispersed teams across the world, a

    robust web-based solution with geographically dis-

  • persed reach was required. They required a solution

    that had an interactive, collaborative business struc-

    K. Kumar, I. Becerra-Fernandez / Decisionture built into the software that would coordinate

    work, and by extension, the people performing the

    work. By doing so, work could be put into context:

    exactly what work is required? Who is asking for the

    work to be performed (i.e., who is in the role of the

    Customer)? Who has to perform the work (i.e., who is

    in the role of Performer)? When is the work due?

    What are the conditions for satisfaction? Which of

    these have been met and which remain unmet? More-

    over, International wanted a system that could help

    achieve continuous process improvement by creating

    a database of past experiences. This, in turn, would

    help improve customer and supplier experiences with

    the system, thereby improving longer term customer

    satisfaction and customer relationships.

    The PCM process has for key phases:

    Change requestwhere all ideas for product

    change and innovation are created and debated.

    Initiated in response to a market opportunity or

    customer demand.

    Change proposalelectronically routes the ideas

    for review, approval, or rejection and results in a

    detailed functional specification for a new or

    improved component or system.

    Change developmentinvolves establishing and

    staffing an engineering team, as well as providing

    the authorization of work and the full design,

    development and engineering release of the change

    order. This phase requires intense negotiation.

    Change implementationwhere new product fea-

    tures or enhancements are put into full production.

    Fig. 5a illustrates the PCM process.

    To better support the PCM process, International

    developed the PCM system, a collaborative commerce

    application centered on managing and coordinating

    the negotiations that constitute this process: requests,

    collaborative agreements and commitments, and

    approvals.11 Its objective was to coordinate work

    between its geographically dispersed sales teams, sup-

    pliers, and customers in an integrated supply chain.

    11 The PCM system is based on applying the ActionWorksR

    Business Interaction Model embodied in the ActionWorksR Metrosoftware.The streamlined PCM process is described in terms of

    the loops that form the core of the system design and

    implementation.

    For example, a program manager from one of the

    product centers acts as the customer in the Business

    Interaction Model. Typically, a program manager

    responds to customer requirements and market infor-

    mation by generating the work authorization (or cus-

    tomer request in the Business Interaction Model) and

    initiates the preparation stage of the model. The lead

    engineer, acting as the performer in the Business

    Interaction Model, receives the work authorization.

    The work authorization can included one to many

    changes or work packages.

    Once those work packages have been reviewed,

    they are distributed to various internal teams and

    outside suppliers (performers) according to a set of

    rules enabled by the principles of the conversation for

    action (Fig. 2) and the Business Interaction Model

    (Fig. 3). The Model allows work to be performed by

    either parallel or sequential processes. The program

    manager is not obliged to manually advance work

    since the system does this automatically. Regardless

    of the group receiving the work package, there is

    opportunity for negotiation and information sharing

    between participants. These interactions are unlimited

    and can be passed back and forth until both parties are

    satisfied with the results.

    One of the key benefits of using the Model is

    that no step in a process is considered complete

    until it is done according to the satisfaction of the

    customer for that particular step. In other words, a

    business process designed using the Business Inter-

    action Model becomes a series of loops, represent-

    ing interactions between parties involved at different

    stages of the overall process. When a step is done

    correctly a loop is closed, and the process con-

    tinues. Using this model ensures that each constitu-

    ent step of a job is done accurately and to closure,

    leading to a desired end-result of accountability and

    quality assurance.

    In PCMS, the business process is launched by a

    request from the program manager. The four stages in

    the Business Interaction Model as handled by PCMS

    are outlined below:

    Preparation: The program manager issues a work

    Support Systems 43 (2007) 584606 599authorization through PCMS to the engineering team.

    As a result, one or many work packages are associated

  • cisionwith the work authorization. Once the lead engineer

    receives the work authorization, the negotiation pro-

    cess begins.

    Negotiation: The negotiation process begins with

    the program manager via the PCM system, to clarify

    the conditions of satisfaction, which are made expli-

    cit via the PCM system, including the work to be

    performed and the date by which it is to be com-

    pleted. Only then does the process move into the

    performance stage.

    Fig. 5b represents the work authorization at this

    stage of the process. The negotiation is a critical step

    in Internationals PCMS. Having a clear understand-

    ing of the engineering specifications has reduced the

    overall rework of Internationals product changes by

    33%, representing significant cost savings.

    Performance: Once the work authorization has

    been agreed upon, the change request moves into

    the performance phase of the process. Each of the

    engineers involved in the change request work on

    their respective portions of the change. Each work

    authorization can generate a set of cascading commit-

    ments (both in terms of specifications and deadline)

    within engineering and between engineering and its

    suppliers. Moreover, the engineers and suppliers are

    often geographically dispersed thereby creating the

    need for managing commitments over distance. The

    production processes to develop and deliver a work

    authorization is complicated, requiring frequent inter-

    action between various participants: engineers, pro-

    gram managers, and external suppliers. Traditionally,

    this interaction consisting of customer requirements

    specification, clarification, negotiation, and approval,

    took place through voice (telephone and meetings),

    paper, faxes, and e-mails. In the past, often product

    changes would sit idle for days or even weeks, lost in

    the system. Together, the demand and supply side

    complexities led to problems of unmet specifications,

    inadequate designs, and missed deadlines. Informally

    documented requests and casually tracked commit-

    ments lead to misunderstandings, frustrations, and

    potential conflicts between various parties thereby

    often leading to damaged relationships and trust.

    Moreover, combined, these issues, together with the

    frequent breakdowns in the process, made cycle times

    unnecessarily long and hampered productivity.

    K. Kumar, I. Becerra-Fernandez / De600Gathering information about the status of the change

    requests would consume around 16 h of the engineer-ing staff week. The PCM system provides complete

    visibility into the process to all of those involved. In the

    words of Lynn Wolfe, head of one engineering group:

    bIn the past I would receive a phone call demanding toknow the status of a work package. It would take 20

    minutes to find the needed information, and I would

    often spend 23 hours on the phone per day. The new

    system cuts search time from 20 minutes to under 1

    minute, and I get fewer calls, because all the parties

    involved in the process have access to PCMS.Q

    Computer support for such a system, in addition to

    keeping track of the requests and commitments, pro-

    vides real-time visibility to all participants throughout

    the supply chain. Everyone involved is able to track

    the status of a work authorization and its related

    commitments and their contingent dependencies to

    deliver particular elements of the application at spe-

    cific times.

    Acceptance: Once a work package is complete, the

    lead engineer sends the change to the engineering

    release integrity group (ERIG), to approve or reject

    the work performed. In addition, a release review

    board, consisting of representatives from engineering,

    product centers, manufacturing teams, and program

    teams further validate the work package completion.

    The review board utilizes a process named the Gaunt-

    let, for reviewing and approving the work packages.

    Fig. 5c describes the Gauntlet.

    The Gauntlet is used to determine compliance with

    the conditions of satisfaction: checks the design intent,

    verifies design standards compliance, enforces data

    management procedures, interfaces to a myriad of

    computer systems, performs automatic auditing of

    databases and CAD models and coordinates antici-

    pated changes at plants prior to the formal introduc-

    tion of the product change. Successful compliance

    with the conditions of satisfaction will result in the

    release of the work package. Fig. 5d illustrates the

    queue of work packages pending release by the

    review board.

    Work packages that do not conform to the condi-

    tions of satisfaction are returned to the performance

    phase. This closed loop method that requires negotia-

    tion of, and commitment to, the conditions of satisfac-

    tion, and customer acceptance of all work has enabled

    Support Systems 43 (2007) 584606International to reduce the rework in its product

    changes by 33%. Essentially, only approval by the

  • Review Board followed by actual review by the pro-

    gram manager, who acts as the process customer, can

    accept the work package and close the loop.

    7. Summary and conclusions

    They key contribution of this paper is the descrip-

    tion of an information technology-based process for

    building relationships and trust. This understanding is

    important in the current environment where business

    is being conducted with trading partners across global

    distances, often with no previous history of interac-

    tion. Therefore, it becomes important these transac-

    tions are supported with interaction technology, which

    enables collaborative commerce at a distance and

    helps to build trust and relationships over time. The

    process of building relationships is commonly con-

    sidered to be a closed loop cycle between a request

    and its satisfaction. Each successful completion of the

    cycle contributes to enhanced levels of trust between

    the partners. Moreover through dialog, the parties

    develop joint awareness of their own and others

    motivations, investments, and perceived uncertainties

    in the business deal. This mutual sense making

    between the parties leads to greater understanding

    about each other and about their own role in the

    transaction. If successful, the increased transparency

    leads to improved relationships and trust.

    The closed loop requestacceptance cycle and the

    joint sense making are based upon conversations

    between the transacting parties. These conversations

    consist of networks of commitments embodied in

    speech acts. However, as the interactions become

    a

    F I L T E R

    CHANGE REQUEST

    (APCN MPCR, Other)

    F I L T E R

    CHANGE PROPOSAL

    (PCBL,MPCR, Concept Shts

    F I L T E R

    CHANGE DEVELOPMENT

    (Plant Processing)

    F I L T E R

    CHANGE IMPLEMENTATION

    (Plant Processing)

    CONCERNS(DCR, e-sheet,

    PIF, FSR)

    IDEAS (Product Plans,

    COMPASS)

    External Influences (Legislated,

    Forced, Vendor)

    K. Kumar, I. Becerra-Fernandez / Decision Support Systems 43 (2007) 584606 601bFig. 5. (a) Product change management process. (b) Work authorization. (c) The Gauntlet process. (d) Release review board queue.

  • heckeeview&

    ECMrificati

    sign Meview&

    ECM

    cisioncCR

    Ve

    DeR

    K. Kumar, I. Becerra-Fernandez / De602numerous, dynamic, and complex, keeping track of the

    commitments by informal manual means becomes

    difficult and error prone. This leads to unmet commit-

    ments and consequent conflict between the transacting

    parties. On the other hand, meticulous management of

    the speech acts and commitments through informa-

    tion technology can lead to explicit satisfaction of

    clearly stated conditions for satisfaction, thereby clos-

    ing the loop. Computer-based commitment manage-

    d

    Start Release Review

    Release Pre-

    Certification Audit Program

    (T Status)

    Verificati

    ERIG Review

    ReleasReview

    BoardReview

    Fig. 5 (contir

    on

    gr

    Support Systems 43 (2007) 584606ment systems, like ActionWorksR Metro helpmanage complex networks of commitments. By mak-

    ing the speech acts and their structure explicit, they

    further help in developing a keener mutual under-

    standing between the transacting parties. Together,

    the ICT support for the two effects, closed loops

    and joint sense making, improves the relationship

    between the two parties and increases the levels of

    trust.

    Approved&

    PendingRelease

    Release TimeAudit Program

    (E Status)

    on

    e

    nued).

  • In conversations for action and in ActionWorksRMetro, we have the beginnings of the principles and a

    prototype tool for building commitments, relation-

    ships, and trust. However, this advantage does not

    come without price. First, there is always the danger

    that the computer-based tools may be used as instru-

    ments of control, and not as intended, for supporting

    variety speech acts [[43], p. 58]. In his later work, he

    classified all speech acts as embodying five funda-

    K. Kumar, I. Becerra-Fernandez / Decisionhuman awareness of networks of commitments. Thus,

    a control-oriented supervisor could presumably use the

    tool as a device for tracking his or her subordinates

    unmet closed loop performance, and use it to penalize

    or reward the performers. Such use would be a perver-

    sion of the spirit of the tool. Second, by insisting on

    making the speech acts and their relationships explicit

    very early in the relationship, it is likely that the parties

    in the transaction may not have opportunity to discover

    common ground and may prematurely decide not to

    pursue the relationship. Some ambiguity in early

    stages of interaction is useful in achieving buy-in

    from diverse stakeholders (Yates 1985). Initial vague-

    ness gives each party an opportunity to perceive their

    own interests reflected in the ambiguous statements

    thus giving them a rationale to participate in the joint

    sense-making process.12 The third price could be

    considered as both a cost and a blessing. Computers,

    unlike humans, have a long memory unbiased by

    recent events. Computers never forget, and cannot

    forgive. On the other hand, Solomon and Flores sug-

    gest that trust can exist even in the light of past

    breakdowns. Reliance on a computer-based commit-

    ment management system may perpetuate the memory

    of broken closed loop cycles far beyond their time.

    Ultimately, however, trust and relationships are

    social processes. The underlying processes of joint

    sense making and the development of mutual under-

    standing are basically results of human interaction. It

    is human intentions, motivations, and emotions that

    drive these processes, not the technology based tool.

    The most such tools can hope to do is to make the

    12 The need for specificity is likely to be a culture-specific trait. In

    task-oriented cultures such as the United States, early specificity in

    contracting is desired for. However, with increasing awareness of

    uncertainty and complexity, scholars and practitioners are beginning

    to realize that contracts are often not complete, and early specificity

    may lead the interacting parties onto a non-agile, inflexible trajec-tory. Consequently, parties are beginning to look at relational con-

    tracts. However, relational contracts in turn require a degree of trust.mental illocutionary points that are supposed to cover

    all speech acts or performative utterances. These are:

    Assertives: Commit the speaker to something being

    the caseto the truth of the expressed proposition.

    For example, the assertive bwe have sufficientinventory of orangutans at this time at this timeQcommits the speaker to producing a sufficient

    inventory of orangutans when challenged.

    Directives: Attempt to get the hearer to do some-

    thing. These include both questions (which direct

    the hearer to make an assertive utterance) and

    commands (which attempt the hearer to carry out

    some linguistic or non-linguistic act). Example,

    bPlease deliver 300 orangutans by noon tomorrow.QCommissives: Commit the speaker to some futureparticipants in the interaction articulate their commit-

    ments, help them be aware of them, and help them

    keep track of them. Thus, for what purposes a tool is

    used, and how it is used depend, to a very large extent,

    on the wisdom of the user. Like any other tool, the use

    of such computer-based interaction tools needs to be

    tempered with and understood in the context of

    human intentions and wisdom. As long as the use of

    these tools supports human processes, and is not

    intended as a replacement for them, they are likely

    to be useful tools for building commitment and trust.

    If on the other hand, interacting parties rely on them

    as alternatives for commitment, relationships, and

    trust, they could be disappointed. Tools, however

    sophisticated, cannot have wisdom built into them.

    It is up to their users, humans, to supply that.

    Appendix A. A brief introduction to speech acts

    Adapted fromWinograd and Flores [[43], p. 5860].

    Speech acts, originally conceptualized as performa-

    tives by the philosopher J.L. Austin [2], are a class of

    utterances that do not refer to the state of the world,

    but in themselves constitute acts such as promising,

    threatening, or declaring.

    Austins student Searle [34,35] formalized the

    structure of the felicity conditions associated with a

    Support Systems 43 (2007) 584606 603course of action. Example bI will deliver 300 oran-gutans by noon tomorrow.Q

  • really not a promise unless there is a mutually

    understood (explicitly or implicitly) time for satis-

    cisionExpressives: Express a psychological state of

    affairs such as apologizing or praising, for example

    bYou did well in delivering the orangutans.QDeclaration: Bring about correspondence between

    the prepositional content of the speech act and

    reality; for example bNow I pronounce you manand wife.Q Or bI hereby confer upon you the degreeof Doctor of PhilosophyQ or bI hereby appoint youas our designated supplier of orangutans.Q

    Searle distinguishes between the illocutionary

    point of an utterance, its illocutionary force, and its

    prepositional content. Illocutionary point is one of the

    five categories above. Two speech acts may differ in

    their illocutionary force while having the same illocu-

    tionary point (bWill it be possible for you to do it bytomorrow?Q vs. bDo it by tomorrow or else!!Q) Thefact that an utterance involves a proposition about

    some topic, such as the delivery of 300 orangutans

    at a particular time, is prepositional content.

    The essential importance of illocutionary point is

    the specification of meaning in terms of patterns of

    commitment entered into by the speaker and the

    hearer by the virtue of taking part in the conversation.

    The taxonomy classifies the possibilities for what a

    speaker can do with an utterance. Different cultures

    may have different uniq