2nd Congress results

Based on the Federal Information Management (FIM) method, the project group is working on the AI-supported analysis of action bases (laws, standards, guidelines, etc.). The information is to be prepared according to the categories of actors involved, required data/information and necessary actions (actions) so that in a next step it can be converted into so-called master information (process models and data field schemas). This master information then maps the operations described in the action bases (eg: "create a certificate") in the form of technical standards (BPMN 2.0 and XML) and can be used for the digitization of application processes up to the provision of services. In order to facilitate this manual and very labor-intensive process, various approaches such as deep generative, deep discriminatory and rule-based procedures are being developed and tested.

The information obtained through the standards analysis will serve as the basis for the intended implementation of digitalized processing processes for requested services in the simpLEX project. The master information already mentioned is to be prepared using an extended concept of the FIM method for the direct implementation of the process model in a technical system of a local target authority. Specifically, the rough process steps are to be replaced by standardized task modules so that a greater level of detail can be achieved. In addition, the openDVA working group is working together with the company betterlaw knowledgeTools to expand the FIM standard to include the decision mapping component and integrate it into the process model. The prepared master information – or local information at this point – is to be imported into an open low-code environment so that digital application submission and processing can be implemented without complex programming work and with improved integration of the specialist knowledge. The possible connection of various basic services and platforms, such as payment components and authentication mechanisms, is also being investigated and tested. The resulting data exchange is researched under the term semantic interoperability.

A large amount of information is collected and processed in both Canareno and simpLEX. This is where the KollOM-FIT project comes in. In this project, the information collected is to be processed and linked in such a way that it is not only available within the openDVA projects, but also in other contexts, directly and machine-readable . Technologically, this work is based on a so-called knowledge graph, via which information from a wide variety of sources can be managed in a structured manner and semantically related to one another. This rich knowledge base can then be processed and made accessible to a wide range of administrative organizations, citizens, political actors, etc. The flexible way in which data is stored means that all actors can be provided with an application that is suitable for them. However, this approach goes beyond human users and also enables further processing in applications such as the automatic generation of forms, as demonstrated in simpLEX. The shared database ensures that all actors, regardless of the specific processing, always have access to the same information, which only needs to be managed in one central location.

The conference ended with a short reflection session with the conference participants. It became clear once again how important events like this are for joint and mutual learning from one another – especially in the cooperation between science, business and public administration. In addition, questions were raised about further research, e.g. about in-depth analysis of LLMs with regard to the assumption of individual tasks in the context of administration or about further formats with a stronger involvement of the administration side. The next steps are therefore initially a stronger networking of those involved and those interested for an intensified, continuous exchange, also with the perspective of further research projects.