AI orchestrates the S/4 transformation

Migration to SAP S/4 via the application layer is one of the outstanding advantages of the JiVS IMP platform. The more than 3,000 business objects that JiVS IMP supports as standard accelerate the migration to the new software generation from Walldorf enormously and generally reduce the associated effort by half. This applies even to the largest and most complex migration projects. However, in these projects, it is not the 95 percent of business objects that are supported in the standard that ultimately generate the greatest effort, but the remaining 5 percent that have to be created additionally. Only with their help can the migration be fully completed - in other words, the required data can also be transferred from non-SAP and individual systems to SAP S/4 Hana.

Business Object Proposer

Companies are aware of the effort involved. Many postpone it, push it aside or avoid it altogether. No wonder, then, that in historically grown and therefore extremely heterogeneous IT environments, more than 50 percent of existing systems are non-SAP solutions - both those from third-party providers and individual solutions. The problem: over time, these systems fall out of maintenance and the risk to data security increases dramatically. For legal and business reasons, companies must continue to ensure access to the data stored in the legacy systems, sometimes for several decades. Added to this is the associated administrative effort. In addition, over time, knowledge of how these systems work is increasingly being lost, a challenge that is becoming ever more acute in view of demographic change and the associated labor shortage.

Effort, knowledge, security - the risks of an incomplete migration are too great to keep them under control in the long term. Against this backdrop, companies need a solution that massively accelerates the design of individual business objects. To this end, DMI has used artificial intelligence to supplement and expand the popular JiVS Business Object Designer (JOD) with the new Business Object Proposer (BOP). The tedious and time-consuming tasks of finding the correct and complete data in the right tables and linking them together correctly are now taken over by the intelligent Business Object Proposer. This has the potential to halve the development time for business objects.

BOP is based on a specially trained large open source language model (LLM) and rule-based algorithms. To ensure full data sovereignty and the necessary data protection, customers can install and use BOP including the language model on-premises. Another advantage: the requirement to migrate data and documents from legacy systems to SAP S/4 Hana comes from the specialist departments. However, they are not IT or database experts. In order for their wishes to be implemented as efficiently as possible, experts are needed who understand exactly what the specialist departments mean and how the requirements need to be implemented technically.

With BOP, companies need less expert knowledge to meet the requirements of the specialist departments. The existing experts can complete more tasks in less time and concentrate primarily on checking the BOP analyses and proposals. Other teams that do not have expert knowledge of the various legacy systems can use BOP to help create the missing business objects and then migrate the data. This makes project participants orders of magnitude more efficient and productive. Further advantages: Thanks to BOP and the subsequent data migration and secure long-term storage, companies can completely decommission all legacy systems, SAP and non-SAP systems alike, as part of the transformation and migration to SAP S/4 Hana. This usually saves 80 percent of operating costs. At the same time, the right to be forgotten obliges companies to ensure cross-system transparency.

Personal Data Identification

European and Swiss data protection legislation has provided for it for years: the right of employees and customers to be „forgotten“. A simple rule, but one that still drives IT experts to despair. Is a person's data „only“ stored in SAP systems or also in the many different peripheral systems, whether from other manufacturers or in solutions developed in-house? Are there several people with the same name and which of them is the one you are looking for? Is the same person listed under different keys, for example because the employee has worked at the same company several times on their career path?

Heterogeneous landscapes

Hand on heart: Many companies, at least those with historically grown and heterogeneous IT landscapes, are likely to recognize themselves in these questions and can only partially answer them at the touch of a button. And even if they clearly knew in which systems, tables and fields the personal data they were looking for was stored, they would have to delete it specifically in each individual system.

More and more companies are working on making personal data uniquely identifiable across systems. As part of these efforts, individuals are given a unique identification number. These are linked to the unique keys of the various systems in which the data is stored. On this basis, a rule can be created as part of retention management (RM) to delete all data associated with a person from a central location, taking into account all system-specific keys - even if the deletion process then takes place in 27 systems, for example. As correct as this approach is, it is very much geared towards the future. This means that the advantage of centralized and automated management does not pay off retroactively, for example when deleting personal data of former employees or customers who have moved to the competition.

So it's no wonder that many companies are still unable to meet their obligations to delete personal data, even after the European General Data Protection Regulation and the current Swiss Data Protection Act came into force!

Companies that work with the JiVS IMP platform, on the other hand, can master these challenges all at once and from a central location. This is made possible by the new Personal Data Identification (PDI) functionality based on artificial intelligence. It searches all systems, especially those for which there is no or a poorly maintained metadata repository, and classifies tables, columns and fields. It then presents the results of its analysis and classification for review. This gives companies an overview of where personal data is located and they only need to fine-tune the proposed results. This reduces the search effort to a minimum.

Best of all, companies can use PDI to create an RM rule and the associated business object - soon also with the help of the intelligent Business Object Proposer - which will enable them to automatically delete personal data, including the associated documents. In this way, they can ensure that all personal data whose legally prescribed retention period has expired is automatically deleted reliably and punctually at the end of the period. In addition, they can comply with authorized deletion requests outside of these deadlines at the touch of a button from a central location, without having to search for the relevant data and touch the affected systems individually. It is precisely this central location that is required by law, which companies must apply to all systems. DMI has created this location with PDI.

In addition to the question of where personal data is located, the same challenge arises in mature IT landscapes for any type of information - be it documents, reports or historical business data.

„JiVS IMP, I have a problem“

The larger the company, the more heterogeneous the system landscape. The latest product generations such as SAP S/4 Hana exist alongside older SAP systems, but also alongside ERP and other solutions from third-party providers together with individual solutions that may not have been developed further for years. All contain data, documents and reports that still have value for the company many years after they have been created and even archived. However, what is missing, and is becoming increasingly missing over time, is in many cases the knowledge of where to find the report, document or data entry you are looking for. The situation increasingly resembles an archive without an archivist.

The teams in the various specialist departments of companies react almost allergically to the possibility of no longer being able to call up a document or report that is fifteen years or older, even if this only happens once every ten years. And yet it is this reaction and expectation on the part of the specialist departments that regularly causes major problems for their colleagues in IT - not least in transformation and migration projects. While IT wants to transfer as little legacy data as possible to new product generations such as SAP S/4 Hana and would prefer to decommission all legacy systems, it regularly comes into conflict with the interests of business users for the reasons mentioned above. The time, personnel and financial outlay for the transformation increases and the level of expertise that companies need to maintain to provide legacy information remains high. A situation that is becoming increasingly difficult to manage, especially in times of a worsening skills shortage and rising cost and competitive pressure.

Chatbot functionality

At least since the meteoric rise of generative artificial intelligence, it should be clear to practically everyone that searches in a wide variety of systems are a prime discipline for so-called large language models. For this reason, DMI has trained an AI for this use case and integrated it into JiVS IMP as a chatbot functionality.

This makes it possible for employees in a company's finance department, for example, to search for and find data and documents that have been historicized on the JiVS IMP platform for legally compliant long-term storage simply by using natural language input: „JiVS IMP, I have a problem. I need to collect the open items for customer 1020 and the 2019 financial year. Unfortunately, I am not familiar with the individual development in which the documents were created. Could you please find them?“ In response, the JiVS-IMP platform chatbot would display the documents searched for, sorted by hit probability, and ask the user to confirm that it is the information they are looking for.

But the dialog doesn't have to end there. Additional information such as all documents or emails linked to the receipt can also be searched for, found and displayed using natural language input. All of this is possible without users having to be familiar with the source system. The chatbot from JiVS IMP increases ease of use and saves time, while the higher productivity and efficiency of employees saves valuable human resources from the company's perspective. In future, companies will also benefit from the chatbot in connection with the intelligent Business Project Proposer - not least during the transformation and migration phase when switching to SAP S/4 Hana. For example, if the migrator knows the name of a legacy business object but not the associated source system, a question to the chatbot in natural language will soon be enough to find the data and have a new business object created.

However, the biggest advantage of the chatbot is that it helps to reduce the fear of contact and, so to speak, loss in transformation and migration projects on the part of the specialist departments. Once the specialist departments realize how easy it is to access historicized information, their resistance to the comprehensive historicization of legacy information on the JiVS-IMP platform diminishes. The increased user acceptance - one of the most important success factors in extensive and complex IT projects such as the transformation to SAP S/4 Hana - reduces the amount of legacy information to be migrated to a minimum, usually cuts the transformation and migration effort in half and allows not only the legacy systems from SAP, but all legacy systems to be decommissioned. This reduces the original operating costs by 80 percent or more. At the same time, the total cost of ownership of SAP S/4 Hana can be reduced by an estimated 25 percent because companies have to store significantly less data in the live system and can therefore save on expensive main memory. But even if access to information is secured, the real effort remains: the structured transfer of this data into the new system world.

Anyone can transform

Transformation projects are a mammoth task for IT departments. However, they also represent a huge additional burden for specialist departments. They are the ones who have to coordinate with IT which data, documents and reports they believe must be available in the live system after the switch to SAP S/4 Hana. And they are the ones who, alongside their actual work for IT, have to write the necessary specifications so that their colleagues can create the appropriate transformation rules.

To enable perfect results in the shortest possible time during the transformation phase, DMI has developed the new Low Code Transformation (LCT) functionality. It is based on a specially trained AI model that can be used to formulate transformation rules in natural language. LCT automatically converts the input into MS SQL statements, one of the most widely used coding standards in the world. The advantage: MS SQL has not only always been the JiVS-IMP programming language. In fact, it is supported by the vast majority of AI models due to its close proximity to natural language. The greatest benefit of LCT is that it saves valuable human resources in both IT and the specialist departments. The functionality reduces the burden on both in transformation and migration projects. LCT also gives companies more flexibility when putting together the teams required to create the transformation rules. With the help of this new functionality, transformation experts can train less specialized colleagues in a short period of time. Thanks to LCT, they can focus not only on the results of their own work, but also on reviewing the work results of their team colleagues. All this saves time without compromising the quality of the results. 

JiVS creates an end-to-end AI approach across all phases of the transformation: from the design of individual business objects to compliance and user acceptance through to rule definition and data quality. This shifts the focus from system migration to the strategic reorganization of the data landscape - with the realistic prospect of consistently shutting down legacy systems, sustainably reducing operating costs and starting with clean data in the S/4 world.

AI improves data

And there is a magic in every beginning ... Not everyone is likely to think of Hermann Hesse's poetic words in the context of transformation and migration projects. Despite all the complexity and effort involved, however, such projects offer a unique opportunity that would indeed have something magical about it: optimizing data quality and eliminating the sins of the past in data maintenance. This is especially true for master data. The more accurate it is, the greater its benefit in the new system. However, many companies shy away from the effort involved. Although they are aware of the benefits of correct data, they believe that these benefits do not justify the additional work involved during the transformation. With the JiVS IMP platform, this is no longer the case. JiVS IMP automatically sorts out duplicates and thus already reduces possible errors by 75 percent or more. Together with the usual reduction of 90 percent to 95 percent of transaction data and 50 percent of master data required for the migration to SAP S/4 Hana, the effort required to optimize data quality is massively reduced.

But it gets even better: with the help of the new AI-based Data Quality Improvement (DQI) functionality, the remaining effort required for data quality assurance can be significantly reduced - by half or more, according to DMI's experience.

DQI also shows its strengths when it comes to completing data records. An example of this would be missing entries for gender in the salutation of customers or suppliers. To this end, the intelligent functionality analyzes the first names in the data records, taking into account country-specific characteristics that it has learned during training. One example would be the name Andrea, which is female in German-speaking countries but male in Greece.

But DQI can do even more: if no external sources are available, the underlying AI attempts to recognize patterns in the correct data records in order to make suggestions for missing or incorrect data records based on these patterns. Of course, humans always have the final say in deciding whether the results obtained are satisfactory and can make improvements using system prompts. In this way, man and machine form a team that perfectly combines their respective strengths. So that customers can hit the ground running in the new SAP world with clean data. (Source: DMI)