The use case presented is based on real needs. The project process itself is fictitious and loosely based on projects that have already been carried out. With it, we want to illustrate which steps are necessary to accelerate complex business processes with the help of Systemic & AI-based Information Integration and how information completeness can contribute to reducing escalating process costs. To this end, we will now get to know a very special customer and its environment a little better.
The plant manufacturer W is part of an international conglomerate and produces, among other things, industrial plants which, as capital goods, are very expensive to purchase. Accordingly, they are used for a long time for the production of specific materials. We are talking about a utilisation period of fifteen years on average.
The conventional version of an industrial plant consists of many different assemblies that are integrated into the desired plant according to the customer's requirements. This procedure makes each plant a unique piece. Its price is usually around 95 million euros.
With such sums, the preparation of a valid cost estimate is of enormous importance for the manufacturer, as the final price must not usually differ greatly from the quoted price. For a manufactory production, the quotation calculation represents an enormous risk. At W, the calculation therefore took an average of five months per offer. The delay was reluctantly accepted by the customers due to the market dominance as well as the range of services offered by manufacturer W. The delay was also aggravated by the fact that it was not possible to make a significant difference between the offer and the price.
The situation was aggravated by the fact that an industrial plant is a strategic purchase, which is usually preceded by lengthy decision-making processes within the customer's organisation. The time until productive use therefore urgently needed to be accelerated on the manufacturer's side.
The company's goal was therefore to be able to provide highly complex quotations automatically and thus faster, more efficiently and with reduced risk. The automation of the quotation calculation in the form of a calculator or configurator was obvious. But what information had to flow into it and how long was it usually valid? And: What had to be done and how in order to access this information?
An information assessment looks at all processes of corporate value creation that can contribute to the fulfilment of information needs, i.e. across departments and systems and across all organisational units involved. Since only complete information on a complex of topics makes well-founded entrepreneurial decisions possible, the fundamental questions are: Is it really complete or do I only assume that it is complete? Who guarantees me its information value?
In order to guarantee these conditions, in our example the processes needed for costing, their stakeholders as well as the customer information (input) needed to start each process step must be analysed, understood and their benefits transparently presented. The same applies to the information created during the process step (output) and its effects on the original information content of the input (feedback).
Coupled and especially decoupled information flows of the process chain, in the sense of missing interfaces (media discontinuities) should thus become transparent.
If one also brings into play the data management practised by the client and, in particular, the usually inadequate data quality often found in the departments concerned, it becomes clear that the insights gained through an information assessment can even call established decision-making processes into question.
One thing should be clear: No tool can help here. The tool to be used later, the Information Integration Platform, which takes care of the integration of all identified information, only indirectly fulfils the customer's need and is of secondary importance for the time being. Later, it makes it possible to implement the customer's respective use case quickly, transparently and securely. Maintenance measures and changes can be carried out centrally and have no impact on the databases of the productive systems that interact with the platform.
Back to our case study. After the commissioning of an information assessment, all specialist departments of manufacturer W that were involved in the calculation were informed and requested by the management to support the consultants. This quickly resulted in points of friction between the data owners themselves and with regard to the possible overall solution. This escalated at the beginning of the project when the information content of the respective data silos was doubted by representatives of neighbouring departments. The data owners also disregarded the instructions of the management and stayed away from the kick-off meetings because they feared that their productive systems would be affected during and after the project. This attitude of refusal urgently needed to be resolved.
Note: Basically, it should be noted that the focus of an organisation's IT department is technical. The organisational processes and their associated data concern technical departments and are usually considered a business management issue.
An experienced project manager who had systemic training in change management took on the roles of mediator as well as conflict coach. Using a methodology that was individually tailored to the client, he was able to explain to the data owners within a one-day workshop how the desired result would benefit all parties involved. In the end, all stakeholders were united by the common conviction that the desired time savings would lead to various process optimisations and thus to relief in day-to-day business.
Thus, the creation of an information map for the quotation calculation process could be started.
The data catalogue of the IT department, which already existed in the company, was not to be used, as it provided - according to its purpose - a purely technical view of the existing databases, servers, etc. The data catalogue was not to be used. In contrast to this, an information map contains all information attributes as well as information flows of the company processes necessary for the desired information completeness.
It is important to identify the contents of all information sources (structured and unstructured data as well as mental knowledge) necessary for the respective process in the company and to document their information including their situational classification and dependencies (data interdependencies) in the process. Several iterations are usually necessary to complete the information map, as further sources of information and thus stakeholders not previously considered can be identified in dialogue with different stakeholders.
At the end of a successful information assessment, the information map documents the technical view of all solution-relevant digital and manual value creation processes as well as the formal and informal information flows. In addition, the map provides an annotated overview of objects, attributes and meta-information that are suitable for satisfying the desired information needs.
The information from the departments involved in the quotation process came from digital systems of all ages, which were available for analysis in networked or non-networked form. During the assessment, additional sources of information were identified in the course of the project, while previously used sources were removed from the bidding process due to insufficient data quality. These were both structured data (i.e. from the attributes of any database table) and unstructured data (i.e. from documents or files of any format from the design department, from the warehouse, etc.).
A large part of the relevant information was found in Excel tables and in the "heads" of individual employees. This mental knowledge had been created over the years through media discontinuities within the digital system world and was not documented anywhere. There were also various non-networked computers in production and in the warehouse that still ran under MS-DOS and still had calculation-relevant information.
In addition, internal abbreviations and technical terms had to be broken down and documented. For example, what was called a "DIN standard" in one system was called a "norm" in another, more production-related one.
Within a month, an information map emerged that clarified the following:
A) Due to inconsistencies - caused by the multiple entry of the same information with different time stamps within the departments involved - the manual rework in compiling the data needed for costing caused the greatest time expenditure.
B) It also became clear that errors had crept into every calculation to the disadvantage of the manufacturer due to the above-mentioned deficits.
C) There were also assemblies that were largely identical in terms of their function and thus caused double costs in administration and logistics.
In this way, previously unknown savings opportunities were identified in the assessment phase, which drastically demonstrated the lack of information quality to the client. At this point, the project had already paid off financially for the client. But now the relevant information had to be brought together and made available to the configurator/calculator for further processing.
In our example, the information integration platform serves as an engine that supplies the desired calculator or configurator with complete information and ensures the smooth running of the information interlocking. The use of an information integration platform makes investments in higher-level systems that can serve an extended information requirement unnecessary. Established applications can therefore continue to be used in the departments. Even new application scenarios or information needs can be realised quickly with an information integration platform.
Note: Evaluation criteria for selecting a suitable platform can be roughly derived from the ten stages of information integration. These have already been presented in the following article:
Some solutions, such as Sherlock from Fischer Information Technology AG, already fulfil the technological criteria presented in their entirety. As defined in Step 8 of Information Integration, we generally prefer systems that enable a central database (repository). The project process described in this article should make it clear why.
In order to achieve the desired shortening of the calculation process, the planning of the technical measures for information integration could now take place after the successful identification of all information sources. A tool benchmark with standardised criteria was used to select a suitable platform. This includes an evaluation of the existing functions, the operating models, the implementation effort as well as an economic consideration of the manufacturer.
After the selection, the planning of the operational processes and the configuration of the information integration platform took place. Commissioning began with linking the required information to the information sets needed by the configurator/calculator.
For this purpose, the identified information from the departments involved was automatically transferred into the format of the selected information integration platform via a conversion interface. This was done via a time-scheduled trigger (push principle) in coordination with the respective data owners.
The integration of unstructured data took place AI-based. In the process, documents are broken down into their elements and assigned to their respective context, i.e. the specific attribute of a database table.
Different attributes of the same meaning were also automatically standardised, e.g. “DIN standard” and “Norm” became DIN or manually provided with metadata to clarify logical connections and to enable new links. Furthermore, the consultants created extended mathematical and logical rule sets as well as conversion tables within the platform. The integration of a translation service also enabled the automated translation of defined information into different national languages.
Along the way, new information was created from previously independent information. A simple example: System A provides the length, System B the width. The area = length * width results for the first time from information integration. The information area is the result of automated information generation.
Individual data can thus be decoupled from the limited departmental view and combined for extended information retrieval.
Within a few weeks of implementation, configuration and testing, the plant manufacturer W achieved the desired information completeness. All information relevant for the calculation was available on a daily basis as information sets via an interface. Every change in the data sources updated the content of the respective information set after twelve hours at the latest.
The sets were output via a web interface that depicted the configurator requested by the customer. Since all information was processed by the Information Integration Platform, the web frontend required little intelligence.
The logical linking of the information sets takes place exclusively on the Information Integration Platform - which considerably simplifies the maintainability and flexibility of the content and the frontend.
Thus, the initial development could concentrate on the usability of the user interface.
At the end of the project, the quotation time of manufacturer W could be reduced to one day. The project duration for this use case based on real projects is four months on average. Experience shows that after a first successful project, the presented approach encourages the customer to want to provide different internal or external target groups with a complete information view. Be it to improve further decision-making processes or to accelerate the time-to-market.