WorkBrain: Merging Organizational Memory and

Workflow Management Systems

Christoph WargitschBavarian Research Center for Knowledge-Based Systems (FORWISS)

Am Weichselgarten 7, D-91058 Erlangen-Tennenlohe

Thorsten Wewers, Felix TheisingerUniversity of Erlangen-Nuremberg

Computer Science Research Group BMartensstr. 3, D-91058 Erlangen

Abstract. Despite the enthusiasm, the workflow management idea faces currently, some problems occur when settingup large workflow applications for complex business processes. To solve some of these problems, a combination ofworkflow management concepts and the notion of "organizational memory information systems" is suggested. The basicidea is to create an evolutionary workflow management system using an organizational memory storage componentconsisting of a workflow case base to save the workflow lessons learned and a storage for the general domain knowledgeof an enterprise. The concept and a prototypical implementation of the system are presented. The example workflow weuse to illustrate the system functions is the inquiry/proposal process of a roller bearing manufacturer.

1 IntroductionOrganizational Memory Information Systems (OMIS)can be seen as a tool to support the management ofknowledge in an enterprise. One of the distinct cha-racteristics of an OMIS is to support a temporal inte-gration of the knowledge an enterprise generates[StZw95]. Workflow Management Systems (WMS) arean instrument to support the execution of businessprocesses. They provide the possibility to save know-ledge along the time axis since workflows have an in-trinsic time-logic structure and furthermore mirror thepresent business context of a certain period of time.Therefore, it seems to be a promising approach to con-nect the concepts of Organizational Memory (OM) andWorkflow Management and merge their technical im-plementations – the OMIS and the WMS.

Since business environments are obviously dynamicsystems, OMIS and WMS must be flexible enough forallowing frequent changes. They have to be able toevolve and improve with time. In this paper, we intro-duce an approach for an evolutionary OM-based WMS.It is supposed to help enterprises to get error-free, ef-fective and efficient, IT (information technology)-enabled business processes. Key elements of the WMSare a process-oriented OM storage layer and OMISfunctions which have the purpose to retain know-howthat is generated permanently while running the WMSand ensures progress along a process learning curve.The OM is kept almost automatically updated since itis permanently fed with information from currentlyperformed business processes. The tasks of the OM-based WMS are divided into three parts: increasingmastery of a company’s workflows, continuous im-

provement of the workflow quality, and retention ofprofessional knowledge. These tasks are interdepend-ent. The stepwise mastery of workflows, for example,contributes to an efficiency improvement. In turn, theweakness analysis in the course of the continuous im-provement process could lead to a reduction of work-flow variety - simpler workflow structures result. Theretention of professional and organizational knowledgeis the foundation to accomplish the other tasks since itsaves "the lessons learned" and ensures proceeding onthe learning curve.

2 Workflow ManagementThe Workflow Management Coalition (WfMC), an in-ternational association of WMS producers, consultantsand WMS users, defines workflow management as:"The management of processes through the executionof software whose order of execution is controlled by acomputerized representation of the process"[WfMC94].

The life cycle of a workflow can be divided in differentphases that usually are summarized into two blocks[MoRW96]: The workflow design phase, also calledbuild time, and the workflow execution phase (runtime). The build time contains business process analy-sis, business process modeling, and workflow model-ing. The run time consists of workflow planning,workflow start, workflow execution and control, andworkflow archiving. WMS are considered to be one ofthe most expanding software markets. Analysts predictan annual 47 percent growth rate for the North Ameri-can and European market until 1998 [NN95]. Espe-cially in Germany the market is booming, particularlyfor document-oriented systems [Yaki96]. Though the

enthusiasm in IT divisions of enterprises is given,workflow management faces currently some problemsin "business reality":

Workflow Modeling Problems and Complexity.Many workflow management projects revealed that it isdifficult to obtain a detailed picture of the real proc-esses including all procedures, official and implicitrules [Herr95; Iten96; MüSc96; p. 7; Scot94]. Work-flow users often state they do not recognize "their"workflow although they got involved in the modelingprocess by interviews. Different interpretation of termsand gaps between real acting and the information aboutacting are some of the reasons why organizationalanalyses only provide an incomplete picture of things[Gapp93]. Additionally, and according to this, businessprocesses are often more complex than assumed[Jost96].

Lack of Flexibility. When managing business pro-cesses with the help of WMS the danger to freeze themis acute [AuBo96, p. 9]. The high costs to implement aWMS additionally hampers necessary changes. Com-plementary to the long-term flexibility, it is importantto be able to react properly to exceptions and distur-bances within a single business case. Only very fewcommercial systems are able to do so [Schn96]. In-flexibility and clumsiness often result from the fact thatworkflows are not designed modularly but consist oflarge, cumbersome units.

Loss of Know-how. For technology-based companiesthe retention and quick dissemination of know-how in-creasingly is an important production factor and criti-cal success factor. However, it turned out to be pro-blematic to collect knowledge, store it appropriately,and retrieve it to solve actual problems within work-flows. Additionally, changes of all kind, especially thebreaking up of grown organizational structures and thereduction of management levels in the course of busi-ness process reengineering measures may cause a lossof know-how. Retrograde steps on the learning curveresult. Particularly the middle management turned outto be a key player: technical and organizational know-ledge is often highly concentrated here [Fais96, p. 5].

A combination of WMS and OMIS ideas and the crea-tion of an evolutionary, OM-based WMS should help toremedy at least some of these problems.

3 Organizational Learning and Or-ganizational Memory

Learning organizations are characterized by activelysupporting learning processes of their members andcontinuous self-development [Pedl89]. The objective ofthe learning organization is to recognize and performnecessary change processes itself. Certainly, this insti-tutionalized learning and adaptation culture assumes a"learning culture" which fosters innovations and crea-tivity [BuSc96].

3.1 OL

"Learning organizations ... purposefully ... enhance or-ganizational learning" [Dodg93]. "OrganizationalLearning (OL)" means the learning of an organizationthat leads to performance improvements. OL refers togaining know-how as well as to "lessons learned" in itsnegative and positive sense. Klimecki states that"problem solving networks" that are aligned crosswiseto hierarchies and divisions are the most important car-riers of organizational learning processes. We willshow that the semantics of workflows could be takenfor such networks.

Argyris and Schon define OL briefly as "the detectionand correction of errors" [ArSc78]. Basically, they dis-tinguish two types of OL: "single-loop learning" and"double-loop learning". Single-loop learning is reactiveand means an organization is able to fix errors. Dou-ble-loop learning exceeds this by the ability to questionprocedures, rules, norms, and strategies. Paper andJohnson as well as Nevis et al. emphasize that OL in-cludes personal learning of single members of an or-ganization but goes beyond it [NeDG95; PaJo96].Therefore, OL is more than the sum of individuallearning results [Bala97; Dodg93]. Nevertheless thequestion arises where learning organizations actuallyretain their knowledge. " ... for organizational learningoccur, learning agents’ discoveries, inventions, andevaluations must be embedded in organizational me-mory" [ArSc78].

3.2 OM

The term "Organizational Memory" is not coherentlyused in the literature ("there are as many perspectiveson OM as authors") [Acke96]. Argyris and Schon takeOM merely as a metaphor that is an aid to explain thebehavior of organizations ("organizations do not lite-rally remember"), whereas Sandelands and Stableingrant organizations cognitive capacities and conse-quently take OM for independent [WaUn91]. One ofthe most cited definition was made by Walsh and Ung-son: " ... organizational memory refers to stored infor-mation from an organization’s history that can bebrought to bear on present decisions" [WaUn91]. Thisbasic definition is extended by Stein: OM lead to ahigher effectiveness of an organization, under somecircumstances also to a lower one [Stei95].

3.3 OMIS

Whereas OM is a conceptual term, OMIS try to supportthis concept with information technology. It is difficultto mark off OMIS from conventional information re-trieval systems, databases, and general resources likeorganizational handbooks [AkSt96]. The ambition tocreate with OMIS shared knowledge spaces that spanthe entire knowledge of an organization is not realistic[ScBa92]. The main problem is to interpret the storeddata correctly if the knowledge domain is large. In-stead, tools that support an "OM in the small" have

turned out to be useful in order to perform certain tasksin an organization more effectively [AcMa95].

4 Basic Concepts

4.1 Framework: Double-loop Learning with aWMS

WMS are usually introduced in a company in thecourse of a reorganization project. They are often themost important information system to put organiza-tional innovation into action. Amongothers, the main task of the suggestedevolutionary WMS is to serve as an in-strument for supporting the redesign ofbusiness processes. Reorganization canbe divided into two basic forms: a radicalreengineering (revolutionary) [Dave93;HaCh93] and a continuous improvementprocess (CIP, evolutionary) [Harr91;Imai94]. In order to support the evolu-tionary reorganization approaches, aWMS must be able to grow and maturerather than to be a self-learning system.

The double-loop learning approach ofArgyris and Schon [ArSc78] is taken tobe the basic concept of the evolutionaryoverall system "WMS + organization".The WMS is an instrument for a learn-ing process within the domain "design,configuration, and execution of businessprocesses". Two learning cycles can be elaborated (seefigure 1): an inner learning cycle where a learning bydoing is performed, and an outer learning cycle whichimplies reflexive, observed learning - learning by su-pervision. Both cycles are supported by WMS function-alities. Core element is an OM storage layer with twocategories of contents: In the first place, it containshistorical workflows that are stored with their auditdata in form of cases, secondly domain-specific knowl-edge about the organization, products, and technolo-gies, e.g. an organization database that contains the or-ganizational structure, target measures, business rules,responsibilities etc.

Primary goal of the inner learning cycle is to increa-singly master the planning and execution of workflows.This works as follows: Upon their introduction, work-flows are not modeled completely and in detail. Merelythe rough structure and workflow building blocks arepre-designed. For each business case within this roughstructure the workflow model is configured out of thesebuilding blocks and can be modified during execution(see also sections 4.2 and 6.4). This way we obtainworkflows that are stored in a case base after their exe-cution and serve as templates for new business cases.

By learning by supervision, the outer learning cycleensures a continuous improvement of workflow plan-

ning and execution and contributes to an improvementof the basic conditions. Prerequisite is to recognizeweaknesses and deficiencies. When they are identified,personnel that is responsible for the workflows or theparticipating employees themselves can correct thesystem e.g. by changing target values, business rules,and the activity network. This guarantees a closedcontrol cycle. Additionally, the outer cycle has thefunction to reach strategic goals with the help of theWMS. The major part of this paper refers to the innercycle.

4.2 Workflow Building Blocks and Case-oriented Workflow Configuration

For complex workflow-enabled business processes theauthors introduce a new approach that substitutes theconventional workflow phase model that strictly distin-guishes between build and run time [MoRW96]. Com-plex, partly-structured workflows can barely be mappedto a few basic workflow models. On the other side, it isunrealistic to supply a complete workflow model foreach special case. Firstly, not each exception can bemodeled a priori, secondly, a huge number of variantswould result. Therefore, there exists a notable gap bet-ween the WMS requirements "easy handling" and"standardization" on the one hand and the need to ade-quately map the complex business reality on the otherhand. A solution for this type of problem is a modulari-zation. We do not think of a componentware approachand take the WMS software to pieces rather thanmodularize the workflows itself.

For this, workflows are dismantled horizontally andvertically. Since also users themselves should modifyworkflows, it is necessary to use a plain meta model.Theoretical terms like frames, patterns, class hierar-chies etc. appear deterrent and should not be presentedto the user. A pragmatic, clear structure helps to divideworkflows into "comprehensive chunks".

Organizational Memory

Cases

General Knowledge

PlanningPlanningPlanning

ExecutionExecutionExecution

ArchivingArchivingArchiving

Inner Cycle:"Learning by Examples"

Inner Cycle:"Learning by Examples"

Inner Cycle:"Learning by Examples"

ModificationModificationModification

Outer Cycle: "Learning by Supervision"

Outer Cycle: "Learning by Supervision"

Outer Cycle: "Learning by Supervision"

AnalysisAnalysisAnalysis

Organizational Memory

Cases

General Knowledge

Fig. 1: Double-loop learning with an evolutionary WMS

The following structure is suggested (see figure 3):Workflows consist of building blocks of variousgranularity. At the top level, workflows are dividedinto workflow phases. Each phase contains a sub-workflow, consisting of a network of activities. Theterm "phase" indicates that workflow phases shouldmainly be structured as a sequence. At the finest level,activities consist of elementary actions. For each typeof building block there is a catalog, the items of thecatalogs are either instances of building blocks used inhistorical workflows or generic templates. The various

building blocks have to be con-figured suitably for each busi-ness case (see figure 4). Thecase-oriented workflow con-figuration reduces the com-plexity of workflow modelsdrastically compared to closedworkflow type models. Theconcept of case-oriented work-flow configuration is supportedby a prototypical workflow en-gine we implemented, "FLEX-WARE" [WaWe97]. The basicidea for the engine was tostrictly separate the "bareworkflow engine" and theworkflow specification data.The configuration is supportedby special search and retrievalfunctions and a case-based rea-soning component.

The suggested approach resultsin a modified workflow phasemodel that differs from the

conventional one described in section 2. Business pro-cess analysis and business process modeling in theirexisting forms are substituted by a single analysis andmodeling of the rough workflow structure and the nec-essary workflow building blocks. The exact design of aworkflow is created when it is started – exactly speci-fied according to the requirements of the current busi-ness case. At this point of time it can be tailored to therequirements of the current business case. Thus, the lifecycle phases "workflow specification" and "workflowplanning" merge to the phase "workflow configura-

tion". The conventionalseparation of workflowtype and single workflowinstance fades away. Forthis reason, a modifica-tion cycle is shorteneddrastically and is exe-cuted at the implemen-tation level, not at an ab-stract business modellevel. Furthermore prob-lems, typically occurringwhen transforming busi-ness process models intoworkflow models, va-nish. Nevertheless it isnot possible to drop theanalysis and modelingentirely, since at least"germs" for the maturingof workflows have to bepresent. Starting pointare workflow buildingblocks which are held in

Workflow XYZWorkflow XYZ

Actions/Checklists

Activity 4711Activity 4711

Subworkflow 0815Subworkflow 0815

Activity Catalog

Phase A Phase B Phase C Phase D

Phase Catalog

Fig. 3: FLEXWARE model

Typ Model

Instances

HistoricalBusiness Cases

TemplateBuilding Blocks

Instances

UsualUsualWorkflow ManagementWorkflow Management

Case-orientedCase-orientedWorkflow ManagementWorkflow Management

+

Build Time

Run Time

Fig. 2: Case-oriented workflow configuration of workflow building blocks

libraries and which can be configured freely (see figure2). Currently, we are performing case studies withcustomers of our industry partner COI in order to setup "starting catalogs" which reflect typical buildingblocks for certain types of processes and industries.

The concepts of workflow modularization and case-oriented workflow configuration are important prere-quisites for an OM-based WMS that is flexible enoughto cope with the problems mentioned in section 2.

4.3 OMIS Functions to Support WorkflowManagement Tasks

To effectively support the tasks of an evolutionaryworkflow management with OMIS functions it is nec-

essary to design thestructure and contents ofthe OM storage layer ina process-oriented way.

One aspect of processorientation is to supportforward and backwardcoupling of know-how,i.e. to transport know-how along a workflow inand against the flow di-rection [Stei93, p. 32].Intelligent informationsystems that enable BPRhave to support a"reallocation of know-ledge" in most earlybusiness process stages[Hart96]. For the INAinquiry/proposal process- which we will presentin section 6.1 - thismeans e.g. that an engi-neer (design phase)might have informationfrom the scheduling de-partment (cost estimationphase) about how to de-sign a bearing in order toproduce it economically.

In turn, know-how canbe transported in forwarddirection with a work-flow, e.g. it could bebeneficial to send hintsabout technical problemsalong with the businesscase folder although itmight not contribute to aworkflow output directlybut perhaps has an influ-ence on the productquality. We implementeda couple of functions to

support the forward and backward coupling of know-how (see section 6.4)

Besides these special know-how-coupling functions,there have to be a lot more OMIS functions which fos-ter the processing of workflow knowledge and, in turn,assist the workflow management itself.

The tasks of workflow management can be divided intotwo groups:

1. Professional tasks related to single activities of aworkflow

2. Workflow configuration and control tasks

Execution of activities

Planning of improvement measures

Diagnosis of weaknesses

Archiving of workflows

Monitoring of workflows

Control of workflows

(Re-)configuration of workflows

Input/output of customer information

Input/output of product information

Input of error messages

Input of comments

Finding know-how bearer

Collection of background infos

Retrieval of documents

Picking up improvement suggestions

Illustration of quality deficiencies

Collection of best practices

Collection of poor cases

Classification/storage in case base

Search/illustration of suitable business measures

Early warning of imminent errors

Input/output of organizational information

Inserting/output of business rules

Provision of background infos

Warning of disadvantageous workflow plans

Assistance in case of exceptions

Finding appropriate task performers

Retrieval of suitable activity templates

Retrieval of suitable subworkflow cases

Retrieval of suitable workflow cases

OMIS FunctionsWorkflow Management Tasks

Fig. 4: OMIS functions for workflow management tasks

Both categories have to be assigned to OMIS functions.In figure 4 possibilities to support workflow manage-ment tasks with OMIS functions are shown. We im-plemented most of these functions prototypically (for adescription of some of these see section 6). Functionswhere users have to fill in information actively in orderto obtain a rich and up-to-date OM are critical. At con-sulting companies like McKinsey & Company [Kiel93;Pete92] and the audit firm Arthur Andersen & Com-pany [Quin92], incentive systems have been imple-mented in order to encourage the employees to contrib-ute to the OMIS’ success. A second possibility to over-come this obstacle is to integrate functions which pro-vide information and functions which require the user'sinput.

5 Implementation

5.1 Technical Architecture

The basic system we use is the document and workflowmanagement system BusinessFlow 3.3 of our industrypartner COI GmbH. It is coded in a C++-like language:OEL (Object-oriented Extensible Language), a pro-prietary language designed by COI [COI96]. We madetwo severe modifications of the system’s architecture:First of all, we exchanged the workflow engineFLOWARE against FLEXWARE. It is based ongraphical description files and a database-orientedcontrol mechanism. Secondly, we implemented a web-based user front-end, called WAX (Web-AXessedWorkflow Management). For the latter we added a spe-cial HTTP server to the existing server architecture,also implemented in OEL. The server is tightly coupledwith the DMS/WMS functions and is able to use thewhole functionality of the basic system andFLEXWARE. Its functioning can be described likethis: Each URL request to the server is transformedinto a method call of the DMS/WMS. The system out-puts the desired information – folders, documents,workflow status reports etc. – which the server trans-forms to HTML code, sometimes enriched with Javaapplets, and returns it to the browser. Therefore, themajor part of the original client functionality of Busi-nessFlow and all of the OMIS functions could bemapped to a WWW browser. Each WMS user receivesa start URL leading him to the login page. Accordingto his permissions, e.g. the right to change workflowmodels, specific functionality is offered to him.

For displaying certain graphical information like ac-tivity networks and business charts we implementedJava applets. These and other graphics are provided bya second server, the freeware HTTP server Apache.

5.2 Logical Architecture

In principle, our system consists of three logical layers:the storage layer, the service layer, and the user inter-face. Each component within the layers is described inthe following sections. The architecture is quite similarto most of the client/server-type information systems.

5.2.1 Storage Layer

The storage layer includes several databases and filesystems which can be clustered into three categories.The workflow storage contains a workflow case basewhere completed workflows are stored. Each case con-sists of extended feature vectors which are used for asimilarity search of workflows, stories and remarksabout the workflow. Additionally links to a descriptionfile that contains the graphical representation of theworkflow and its process logic is provided. Further-more, audit data like cycle times, processing times, andcosts which are logged are part of the workflow mem-ory. Beyond this episodic knowledge, the workflowmemory holds the building block catalogs (workflows,sub-workflows, activities, checklists, application systemcalls) and a "general knowledge database" where theproduct groups of INA, the different kinds of engi-neering applications, the market areas, and the respon-sibilities are stored. A rule base is used to decide forwhat combination of these attributes of a business casewhich organizational unit should perform which acti-vity within workflows. The organizational storagemainly holds the organization database (roles, rights,units, and positions) and a bunch of "organizationaldocuments" like ISO 9000 handbooks. The technicalstorage consists of databases that belong to KODAS (amechanical design database), MEDIAS (an electronicproduct catalog), and TADDY (Technical ApplicationDocumentation and Description System) as well as allkinds of technical documents like drawings, norms,and calculation sheets that are stored in the DMS.

5.2.2 Service layer

The service layer components are either directly relatedto storage layer components like KODAS andMEDIAS or serve for different storage elements likethe HTTP server. Some types of application systems areclearly INA-specific or at least typical for an engi-neering environment. The service layer can be dividedinto three basic sections: The control/basic services, theinformation services, and the communication services.Document server, update server, the database server,the short message, and electronic mail mechanism arecomponents which are part of BusinessFlow. KODASand MEDIAS are systems we integrated into our pro-totype. All the other components have been imple-mented in our project and added to the prototype. Ex-perienceFlow is a case-based reasoning (CBR) applica-tion that is used for workflow planning. TADDY is aknow-how database for engineering solutions. EachTADDY document folder contains problem descrip-tions, problem solutions, and the related drawings. Thediscussion forum WIBIS (Workflow Issue-based Infor-mation System) serves as a communication platform todiscuss technical and organizational problems.

The user interface is explained in the next section withthe help of an example workflow.

6 Usage of WorkBrain in the INA-Inquiry/Proposal Process

The OMIS user interface components are part of WAX.With WAX we made almost all "regular" workflow cli-ent functions available in a WWW browser. We areable to: model and configure workflows, monitorworkflows, provide to-do lists, show workflow docu-ment folders, edit and upload documents, and offer ap-plication systems that should be used for certain activi-ties.

It seems to be natural to use a WWW browser as theuser interface for all OMIS functions, too. It integratesall kinds of application systems inside and outside acompany homogeneously, elegantly, and at little costs.This closeness and homogeneity of information and thepossibility to connect informational chunks to storiesenforce the imagination of the user to navigate througha unified knowledge space. Some authors speak of theWWW as global brain [MaBa94]. It is transparent forthe user where the information comes from.

Our system has – of course – a smaller focus. How-ever, since the multiple relationships and links betweenall the information categories resemble kind of a sys-tem of neurons and synapses we called the OMIS partof our prototype "WorkBrain". We tried to visualize theinformation network in form of a "knowledge map"which is clickable. Each icon symbolizes an informa-tion category. The notion of a "workflow" defines thesemantics of the network.

One element of the user interface is the "control panel"of WorkBrain. It provides access to all parts of thestorage layer components except the technical applica-tion systems and databases. Extended search and re-trieval functions are given.

6.1 Process Example

In order to illustrate the concept and the system func-tionality we consider the inquiry/proposal process forspecial bearings of our industry partner, INAWälzlager Schaeffler KG in Herzogenaurach. INAproduces roller bearings, motor elements, and linear-guidance systems for car manufacturers and the ma-

chine tool industry. Standardized catalog bearings aswell as customer-specific bearings are produced. Theset up of a proposal for a customer-specific bearing isorganized in the process phases depicted in figure 5details of the process are described in [MoRW96].

The inquiry/proposal process has a long cycle time andpasses through several hierarchy levels of INA. It con-sists of well- and poorly- structured process parts. Thisspread is challenging for a WMS designer. The pro-cessing of inquiries needs a lot of know-how. The in-volved employees are mostly highly-qualified specia-lists with long-term experience. The high specializa-tion results in a strongly functional orientation and aramified organizational structure. The process costs forcompleting a proposal are about US-$ 3,000. Accor-ding to habits in the machine tool industry, these costsare not billed. The process is important because specialbearings often become catalog bearings and specialsystem solutions that are generated within the processare used as competitive weapons. The process leads toinnovation, and the developments launched by thecustomer ensure that INA stays close to market needs.Even though the catalog business has a higher revenue,the special bearings are of importance because a lot ofcustomers want to get all bearings they need from asingle source. The complexity of the process can simplybe shown by some figures: About 700 people and 300organizational units can be involved in in-quiry/proposal processes. There are more than 100 ele-mentary actions that have to be performed. The cycletime is 55 days on average whereas one business casecan include up to one hundred documents. 2.000 in-quiries have to be processed per year.

The next section describes the WMS functions ofWorkBrain. We take the modified workflow life cyclephases to structure the description.

6.2 Design of Workflow Building Blocks

The first phase in the modified workflow life cycle isthe design of building blocks. Reference process mo-dels or reference function models which mirror thecharacteristics of an industry seem to be a promisingstarting point for setting up building blocks [MoRJ94].The first one – the process reference model – faces the

Register Inquiry

Processing byEngineers

Evaluate Inquiry

Start Proposal Process

Design Product

Make-or-Buy-Decision and Cost Estimation

Complete Proposal

Send Proposal

CustomerCustomer

Field Field ServiceService

Decision

CustomerCustomer

Fig. 5: Phases of the INA inquiry/proposal process

problem to be either too abstract (then the informationneeded to design workflows is too poor) or it is too spe-cific (then what is left over after an adaptation to theown situation tends to zero). The second one – thefunction reference model – does not provide any pro-cess logic. The authors suggest the already explainedapproach of reference building blocks to overcomethese problems: Process logic can be provided at thedesired level (workflow or phase level) and the buil-ding blocks can be continuously shifted between the ab-stract and the concrete level. The building blocks stemfrom three sources:

• a sample library containing real case studies of ourworkflow project of our industry partner COI,

• a template library which contains generalized ele-ments from the case studies,

• and an enterprise-specific database which consistsof parts of the sample database and the case studydatabases, both modified according to the INAprocess requirements.

The only library that is released to be modified is theenterprise-specific library. It contains building blocktemplates for the INA inquiry/proposal process and in-stances (the workflow case base). This library isevolving with time and can be adjusted continuously.New building block variants can easily be created bycopying and changing existing ones. Figure 6 showshow a variant of an activity is created. The father-son-relationships of building blocks are stored in order tobe able to trace its evolution. Each building block canhave the status "active", "obsolete", or "planned". Dueto the outer learning cycle presented in section 4.1 it isnecessary to align the building blocks with higher levelbusiness tactics and strategies. It is a future task of ourwork to systematize the realization of strategic goals atthe operative workflow level while simultaneously re-acting to the recognized weaknesses of the currentworkflows and external influences.

6.3 Workflow Configuration

The start phase of a workflow instance is split up intothree parts: the workflow triggering, the automatedworkflow pre-configuration, and the manual fine con-figuration. An inquiry/proposal process is triggered byan inquiry a customer or a field service employee posts.In WAX both groups have dedicated access to specialforms where they can fill in data like the customername, the desired product category, the application ofthe product etc. By sending this HTML form to theWAX server a workflow folder is created,automatically pre-configured, and the workflow isstarted and routed to the first person handling theworkflow. The pre-configuration already uses know-how saved in the OMIS storage: Depending on a set ofparameters, a default template workflow is combined ofbuilding blocks according to business rules found in apre-configuration rule table. The desired proposal date,e.g., is decisive for the "length" of the workflow: If thisdate is closer than ten days to the present date then inthe workflow phase "cost estimation" a special sub-workflow is assigned (a shortened estimation procedurethat drops the detailed handling of the expectedmanufacturing costs at the involved plants).Additionally to the regular role mechanisms, weimplemented case-dependent roles that may overridedefault roles subject to certain parameter constellations.Parameters are the customer size, the customer region,the industry, the product category, and the applicationcategory. Figure 7 shows the role "abt_leiter_antr"which is chosen if the customer is large, located inSouth/West Germany, the product type is "Axial-Nadelkranz", the application is "Fahrwerk", and theindustry is "Personenfahrzeuge". If this combination isvalid for a workflow, the role is taken for every activityhaving the default role "angestellt_ott".

The first employee dealing with the workflow is the"workflow planner", a person who can be seen as thecounterpart of the classical "work scheduler" formanufacturing operations. He takes a look at the speci-fication of the pre-configured workflow and makeschanges if necessary. The first thing he has to do is tosearch workflow cases which might be logical prede-cessors of the current case. For instance, an order by acustomer is related to an offer he received from INAseveral months ago. At INA people think very much in"business cases". They even sometimes remember thearbitrarily chosen project numbers of past inquiries, theso-called "F-number". Old business cases and their re-lated engineering solutions are often been taken tosolve actual problems. To use existing process know-how for his changes, the workflow planner has twopossibilities: He can regularly search the workflow casebase or use a CBR component. The result of the casesearch is the complete specification of (one or more)historical workflows. Suitable building blocks of thisspecification can be reused by copying, pasting, andadapting it to the current business case. Figure 8 showsthe search mask for historical workflows.

Name

Role

Description

Applications

Condition

Checklist

Status

Fig. 6: Creation of activity variants

Much of the organizational knowledge is implied in theactivities, that are performed in a company, and their

relationship to each other (the workflowphases). In WorkBrain, all workflow ac-tivities can be searched and browsed.Since it is important to get a visual im-pression of a workflow we implemented aJava applet that illustrates the activitynetwork. We use different layers to visu-alize the workflow structure. At the toplevel, each rectangle represents a work-flow phase that contains a sub-workflow.With a mouse click on the phases, theunderlying sub-workflow pops up in adifferent window.

The part organization memory of theWorkBrain control panel provides accessto an organizational database. The data-base can be browsed entirely. Each rela-tion between information categories isrepresented by hyperlinks. For a role,e.g., the corresponding users, activities,and rights are hyperlinked. For the work-flow planer it is easy to get an overviewof the multiple relationships in the or-ganization and manually change activity-role-assignments if necessary.

The second possibility for the workflowplaner to access workflow knowl-edge is the CBR component that iscoupled via DDE (dynamic dataexchange). It uses flat feature vec-tors to find workflow cases. Differ-ent similarity measures can be cho-sen. Results of a CBR search arepast workflows. The process logic,the graphical description file, thedocuments, and the plan data canbe taken to configure a new work-flow or to just take a look at thesolutions and copy the drawings,business data etc. Accordingly,sub-workflows and activities can befound and reused as buildingblocks in a similar manner. Afterthe termination of a workflow, it iscompletely stored with all relatedinformation. Figure 9 shows theCBR search window of the work-flow phase case base and the resultlist. A detailed description of thesimilarity measures and the im-plementation of the CBR compo-nent can be found in [Oed97].

Case-Dependent Role

Standard Role

Activity

Parameter Combination:- Product- Application- Customer

Fig. 7: Modificaton of case-dependent roles

ProcessType

InvolvedUsers

Periodof Time

Customer

Product Type

Application

Project Number

Evaluation

Fig. 8: Workflow search mask

6.4 Workflow Execution and Control

After the workflow planner terminated the workflowfine configuration, he saves it and the actual processingof the inquiry starts. The workflow folder is passedfrom employee to employee and is gradually filled upwith documents like drawings, calculation sheets,forms, price lists, and finally the proposal letter to thecustomer. Due to the process’s complexity and its crea-tive nature, it must be able to change the workflow de-sign "on-the-fly". Every user who has the right tochange workflow specifications can redesign the work-flow with the graphical editor and save the modifiedversion in the engine’s database. Changes can be madefor all future activities and their relationships to eachother. All past activities are blocked.

The rights for workflow modification are coupled withthe regular rights-role administration. It is reasonableto keep the number of employees who are allowed andable to modify workflows as large as possible sincedisturbances and suddenly occurring needs to changethe workflow structure can appear at every state of thebusiness case. If someone else than the affected em-ployee has to be involved in order to change the work-flow, the advantage of "on-the-fly" changes is lost. Onthe other side, it is desirable to keep the number ofworkflow variants as small as possible. Therefore, onlycertain people should have the possibility to declare

workflow instances to be generic workflow templates.Furthermore, the top level workflow that connects theworkflow phases should be "stable", i.e. not be changedfor a period of time longer than one or two orders ofmagnitude of a single workflow's cycle time.

Each activity that has to be performed appears in theto-do lists of all users who possess the related role. If auser chooses an activity with a mouse-click, the activitydisappears from all other to-do lists. A new browserwindow with the "to-do mask" pops up containing theworkflow folder with all documents, some workflowdata, and the trigger button of application systemswhich support the execution of the activity, e.g.TADDY. Additionally, checklists are available whichhave the purpose to guide the employee when per-forming complex activities like the technical assess-ment of an inquiry and also in order to document whathas been done. The items of the checklists can be con-nected via hyperlinks with documents like ISO9000 in-structions or explanations. The results of the work donewith the application systems can be transferred to theworkflow folder. When the activity is completed, theuser closes the to-do mask and the workflow is contin-ued. Figure 10 shows the principle of the workflowprocessing.

Search Mask to FindSimilar Cases

Result List withDescriptions

Fig. 9: CBR search and retrieval

6.5 Knowledge Transfer

In order to transfer knowledge between organizationalmembers in general and to couple know-how forwardsand backwards within a business process, several me-chanisms and components have been implemented:

• The storage of interdependencies between work-flows is established. One of the reproaches againstprocess orientation is that business processes are"cuts through an enterprise’s body" and do notconsider the relationships orthogonal to these cuts[Mert96]. We are at least able to retain the relationof subsequent processes (see section 6.3).

• The cross-sectional transfer of know-how is sup-ported by a special "know-how counting function".Each time a user performs a workflow activity thecounter increases by one. Thereby, it is possible tofind out who is experienced in certain activities.Figure 11 shows the know-how table of SamsaGregor. Especially in a know-how-oriented enter-prise like INA with a high degree of specialization,it is advantageous for a workflow planner to assignthe right person to perform a task. Of course, peo-ple often simply know who has special skills for acertain task, but sometimes this is not the case anda know-how bearer storage is an aid.

• One of the most important design elements of thesystem is the hypermedia user interface. It allowsto present the context of formerly isolated work-flow information according to the philosophy:"everything is a link".

• A similar effect is accomplished by the possibilityto retrieve workflows via document searches. Es-pecially checklists which store also "soft informa-tion" like episodes, problems, etc. contribute tomaking procedural and technical knowledge avail-able – detached from the strict time-logic structureof processes.

• Automated storage of knowledge along with theaudited activities is just one part of the manage-ment of OM. The other element is communicationthat generates and distributes knowledge. The dis-cussion forum "WIBIS" (Workflow Issue-basedInformation System) is based on the idea of issue-based information systems (IBIS).With this forum,it is possible to assign certain topics to roles and

Activity

WorkflowFolder

1. Choose a business case in to-do list

2. Add results to work- flow folder

3. Forward workflow tonext employee

To-do List

To-do Mask

1.1.

2.2.3.3.

Fig. 10: Principle of workflow management with WAX

List of activities thatSamsa Gregor performed,

sorted alphabetically

Fig. 11: Know-how of Samsa Gregor

activities. The roles automatically receive all newmessages in WIBIS. Therefore, the discussion ofproblems can be proliferated in both directions(future and past of a workflow state) within a busi-ness process since e.g. certain roles are only"active" in confined workflow phases. The rolebearer are actively notified about issues exceedingthese limits. Discussion topics can be generatedabout business case problems (technical or com-mercial) or workflow management itself. WIBIScan be "started" anytime. It provides functionalityto place statements and add pros and cons or ex-amples to them (see figure 12). It is also able to sethyperlinks to certain topics or statements thatpoint to electronic documents (filed anywhere, e.g.workflow folders of the WMS or external WWWdocuments) that are of some interest for the topic.

7 Summary and Next StepsThe proposed concept addressed the problems of WMSand OMIS described in chapter 2. The difficulties ofboth systems can be remedied by combining the un-derlying ideas and designing an integrated system withcommon functions - an OM-based evolutionary work-flow system. It represents an OM in the small with thespecific task to support the planning, control, and exe-cution of complex industrial core business processes.Double-loop learning is supported by the concepts"case-oriented workflow configuration" and "hybrid

workflow management", a process-oriented OM storage structure, andspecial OMIS functions. Proceduraland professional know-how is retainedworkflow-aptly in the OM. The exten-sion of a WMS to an OMIS appearspromising since daily work practiceand know-how management are inte-grated. The main focus of the imple-mentation was to integrate the OMISfunctions into the working environ-ment of the employees and to createan "easy-to-use" interface that pres-ents information and functions con-textually.

There are still deficiencies concerningthe outer learning cycles. Work has tobe done to systematize the workflowevaluation, the implementation of im-provement measures, and the relatedOM maintenance. Also a frameworkfor breaking down strategic goals tothe workflow level is still not avail-able.

We currently perform case studies in-vestigating further complex project-like business processes like the in-quiry/proposal process at INA. Theresults will be used to fine-tune theWMS prototype and to implement

further process examples and to create a workflowcomponent library that can be used as a starting pointfor the workflow evolution process when implementingWMS in companies.

Parts of the concepts and solutions of our prototype aretransferred to the running workflow system at INA andto the WMS products of COI: e.g. the FLEXWAREapproach is used for the workflow engine of the nextrelease of BusinessFlow.

8 AcknowledgementsThis project is supported by our industry partners: INAWälzlager Schaeffler KG and Consulting for Officeand Information Management (COI) GmbH and theBavarian Research Center for Knowledge-Based Sys-tems (FORWISS).

Hierarchical DiscussionContributions

Chosen Item

Fig. 12: WIBIS

References

AcMa95 Ackerman, M., Mandel, E., Memory in the Small: An Application to Provide Task-Based Organizational Memory fora Scientific Community, in: Nunamaker, J., Sprague, R. (eds.), Proceedings of the 28th Annual Hawaii InternationalConference on System Sciences, Vol. IV, Los Alamitos 1995, p. 323-332.

AcSt96 Ackerman, M., Stein, E., Organizational Memory and Organizational Memory Information Systems, Tutorial, 29thAnnual Hawaii International Conference on System Sciences, Wailea, Maui 1996.

Acke96 Ackerman, M., Organizational Memory, URL: http://www.ics.edu/ackerman/om.html, 14th December 1996.

ArSc78 Argyris, C., Schon, D., Organizational Learning: A Theory of Action Perspective, Reading 1978.

AuBo96 Aulbur, W., von Bonin, T., Zementiert das System unproduktive Abläufe?, in: Standardsoftware (1996) 7/8, p. 5-9.

Bala97 Balasubramanian, V., Organizational Learning and Information Systems, URL: http//eies.njit.edu/333/orglrn.html,9.5.97.

BuSc96 Bullinger, H., Schäfer, M., Lernende Organisationen, Oracle Welt (1996) 3, p. 4-9.

COI96 Consulting for Office and Information Management (COI) GmbH (eds.), BusinessFlow 3.3 User's Manual, Herzogen-aurach 1996.

Dave93 Davenport, T., Process Innovation, Boston 1993.

Dodg93 Dodgson, M., Organizational Learning: A Review of Some Literatures, Organizational Studies 14 (1993) 3, p. 375-394.

Fais96 Faisst, W., Wissensmanagement in Virtuellen Unternehmen, Working Paper no. 8/1996 of the series "Informationand Communication Systems as Design Elements of Virtual Enterprises", University of Erlangen-Nuremberg, Nur-emberg 1996.

Gapp93 Gappmaier, M., Kooperative Vorgangsbearbeitung - Eine empirische Untersuchung der Arbeitsorganisation, WorkingPaper of the Institute of Computer Information Systems of the University of Linz, Linz 1993.

HaCh93 Hammer, M., Champy, J., Reengineering the Corporation, Boston 1993.

Harr91 Harrington, H., Business Process Improvement, New York 1991.

Hart96 ‘t Hart, M., Matching Intelligent Systems with Business Process Reengineering, Intelligent Systems in Accounting,Finance and Management 5 (1996) 5, p. 41-53.

Herr95 Herrmann, T., Workflow Management Systems: Ensuring Organizational Flexibility by Possibilities of Adaptationand Negotiation, in: Proceedings of the Conference on Organizational Computing Systems (COOCS), Milpitas 1995,p. 83-94.

Imai94 Imai, M., KAIZEN: Der Schlüssel zum Erfolg der Japaner im Wettbewerb, Berlin 1994.

Iten96 Iten, A., Workflow-Management bei Landis&Gyr, in: Österle, H., Vogler, P. (eds.), Praxis des Workflow-Managements, Braunschweig/Wiesbaden 1996, p. 253-270.

Jost96 Jost, K., Pilotprojekterfahrungen, in: Österle, H., Vogler, P. (eds.), Praxis des Workflow-Managements,Braunschweig/Wiesbaden 1996, p. 229-238.

Kiel93 Kiely, T., Learning to Share, CIO (1993) 7, p. 38-44.

MaBa94 Mayer-Kress, G., Barczys, C., The Global Brain as an Emergent Structure from the Worldwide Computing Network,and its Implications for Modeling, in: Tech-Report CCSR-94-22, Center for Complex Systems Research, University ofIllinois at Urbana-Champaign, 1994.

Mert96 Mertens, P., Process focus considered harmful, WIRTSCHAFTSINFORMATIK 38 (1996) 4, p. 446-447.

MoRJ94 Morschheuser, S. et al., Referenzmodell einer integrierten Daten- und Dokumentenverarbeitung im Industriebetrieb,dargestellt am Beispiel der kundenwunschorientierten Anfrage-/Angebotsabwicklung von Maschinenbauun-ternehmen, Working Paper no. 2/1994 of the Department of Information Systems I, University of Erlangen-Nuremberg, Nuremberg 1994.

MoRW96 Morschheuser, S., Raufer, H., Wargitsch, C., Challenges and Solutions of Document and Workflow Management in aManufacturing Enterprise: A Case Study, in: Lynn, M. (ed.), Proceedings of the 29th Annual Hawaii InternationalConference on System Sciences, Vol. V, Los Alamitos 1996, p. 4-13.

MüSc96 Müller, S., Scheidt, M., Geschäftsprozeßoptimierung - mehr als nur ein kurzlebiger Trend!, Industriemanagement 12(1996) 5, p. 6-9.

NeDG95 Nevis, E., DiBella, A., Gould, J., Understanding Organizations as Learning Systems, Sloan Management Review,Winter 1995, p. 73-85.

NN95 N.N., Workflow-Software wird ein Renner, Computerzeitung 26 (1995) 34, p. 17.

Oed97 Oed, J., Entwurf und prototypische Realisierung einer Know-how-Datenbank zur fallabhängigen Workflow-Planungund -Steuerung, Diplomarbeit, University of Erlangen-Nuremberg, Nuremberg 1997.

PaJo96 Paper, D., Johnson, J., A Theoretical Framework Linking Creativity, Empowerment, and Organizational Memory, in:Lynn, M. (ed.), Proceedings of the 29th Annual Hawaii International Conference on System Sciences, Vol. IV, LosAlamitos 1996, p. 20-33.

Pars59 Parsons, T., General Theory in Sociology, in: Merton, R. et al. (eds.), Sociology Today: Problems and Prospects, NewYork 1959, p. 3-37.

Pedl89 Pedler, M. et al., Towards a Learning Company, Management, Education and Development 20 (1989) 1, p. 20-38.

Pete92 Peters, T., Liberation Management, New York 1992.

Quin92 Quinn, J., Intelligent Enterprise: A Knowledge and Service Based Paradigm for Industry, New York 1992.

ScBa92 Schmidt, K., Bannon, L., Taking CSCW Seriously - Supporting Articulation Work, CSCW 1 (1992) 1-2, p. 7-40.

Schn96 Schnitzke, M., Konzeption einer flexiblen Komponente zur Steuerung und Integration in ein Workflow-Produkt,Prediploma Thesis, University of Erlangen-Nuremberg, Nuremberg 1996.

Scot94 Scott-Morgan, P., Die heimlichen Spielregeln: die Macht der ungeschriebenen Gesetze im Unternehmen, Frank-furt/Main 1994.

Stei93 Steinmetz, O., Die Strategie der integrierten Produktentwicklung, Braunschweig 1993.

Stei95 Stein, E., Organizational Memory Review: Concepts and Recommendations for the Management, International Jour-nal of Information Management 15 (1995) 2, p. 17-32.

StZw95 Stein, E., Zwass, V., Actualizing Organizational Memory with Information Systems, Information Systems Research 6(1995) 2, p. 85-117.

WaUn91 Walsh, J., Ungson, G., Organizational Memory, Academy of Management Review 16 (1991) 1, p. 57-91.

WaWW97 Wargitsch, C., Wewers, T., Wirthmüller, A., WWW-Front-Ends for Document and Workflow Management Systems,in: Proceedings of the IT-Vision-Week, CD-ROM, Paderborn 1997.

WaWe97 Wargitsch, C., Wewers, T., FLEXWARE: Fallorientiertes Konfigurieren von komplexen Workflows - Konzepte undImplementierung, Müller, M., Schumann, O., Schumann, S. (eds.), Proceedings of the 11th Workshop 'Planning andConfiguration' on the 4th German Conference "Knowledge-based Systems" (XPS-97), Erlangen 1997, p. 45-55.

WfMC94 Workflow Management Coalition (ed.), Glossary, A Workflow Management Coalition Specification 2/1994, Brussels1994.

Yaki96 Yaki, A., Germany - A high growth market, poised for major change, Document World (1996) 5, p. 14-16.