Dynamic AI infrastructure at RÖDL: Overcoming boundaries

Artificial intelligence (AI) represents a real game changer for RÖDL with a disruptive character in terms of the approach to solving complex issues. With the correct and targeted use of AI, a decisive competitive advantage can be achieved for the entire RÖDL with far-reaching potential and positive effects for our clients. Therefore, we have launched strategically decisive AI initiatives to be significantly in the driver’s seat when it comes to using and realizing AI functionalities for professional services firms.

Mastering challenging problems and understanding highly complex process structures form elementary cornerstones of our daily work. Through the newly initiated era of multi-agent systems with the aid of individual agents of generative AI, the processing of highly complex tasks by AI is of particular importance. In a multi-agent system, several AI agents work together in a coordinated manner within a workflow to specifically take on individual tasks. The division of a higher-level problem into smaller, less complex task packages in a workflow and the targeted formalization for handling each task package within a higher-level problem is crucial. For example, one agent has the role of the executing agent, while the other agent has the role of the supervisor and controls the outputs of the executing agent.

AI knowledge graphs and ontologies

Clear and structured knowledge graphs based on data, context, and ontologies, in relation to the approach to processing a task, form the essential building block. A knowledge graph is a structured, semantically enriched representation of expert knowledge that organizes entities (nodes) and their relations (edges) in a graph model. They serve as the basis for the integration, linking, and querying of heterogeneous data sources, thereby enabling context-aware, explainable, and consistent knowledge utilization in generative AI systems.

In contrast to purely statistical models, a knowledge graph offers explicit semantics that are defined by an ontology. The ontology functions as a formal schema that describes concepts, classes, properties, and logical restrictions. Knowledge graphs are modeled in advance, in particular, by our human expert knowledge from our colleagues in order to provide the AI system with the relevant context information. Through this approach, the achievement of reliable and resilient results from AI systems is demonstrable and a significant added value associated with it can be implemented.

Data Tales research project

In order to further advance this development in the field of AI, we at RÖDL in the field of auditing act as a practical partner in the research project called “Data Tales” funded by the Bavarian State Ministry of Economic Affairs, Regional Development and Energy together with the cooperation partners dab: Daten – Analysen & Beratung GmbH and the Ostbayerische Technische Hochschule (OTH) Regensburg.

The goal is to develop a hybrid AI approach for the automated generation of interactive “Data Stories” in the field of auditing. The focus is on the question of how Large Language Models (LLMs), domain-specific knowledge, autonomous agents, and interactive machine learning can be combined to create narrative data analyses that are both technically valid and understandable and interactive. Through the innovative combination of human-AI interaction and autonomous story generation, intelligent chatbots are used for exploratory data analyses to increase the efficiency and quality of audit processes. The project focuses on the relevance, context, and impact of results, thereby combining data-analytical precision with effective narrative communication in order to comprehensively support or improve the process of an annual audit.

In-house AI infrastructure

In addition to participating in the Data Tales research project, the development of data-sovereign AI is particularly important for RÖDL. Therefore, for the 2026 financial year, an in-house on-premise AI infrastructure with local graphics cards and server structures will be created in close cooperation with our subsidiary, RÖDL IT Operation GmbH. This enables us as RÖDL to have maximum sovereignty over AI-based data processing and a maximum degree of individual customizing in the development of in-house AI systems. To create maximum transparency, RÖDL will also rely on a variety of open source language models of generative AI. These enable targeted and individually tailored fine-tuning to solve a wide range of problems. This guarantees on the one hand complete data control and on the other hand maximum flexibility in solving problems for our clients.

RÖDL AI Hub

The heart of our on-premise AI infrastructure is our RÖDL AI Hub, which has already been in use worldwide for several months. The flexibility of the RÖDL AI Hub is of particularly high added value, as it allows us to take into account the specific requirements from the different business areas in our company. In addition, the RÖDL AI Hub offers the possibility of connecting a wide variety of IT systems, which facilitates the linking of all data sources with AI. This integration promotes comprehensive data analysis and use, which can significantly increase the efficiency and quality of well-founded decision-making processes. In addition, the integration of RPA & No-Code/Low-Code solutions for end-to-end automation of processes is possible, with the RÖDL AI Hub representing the intelligent link. No Code/Low-Code solutions enable applications to be created largely without or with minimal programming effort via visual tools and predefined development components.

Conclusion

The focus of our AI initiative is on the development of AI systems with proven specialist expertise, therefore we attach great importance to the modeling of knowledge graphs and ontologies, which form the basis for our AI functionality.

We as RÖDL are greatly expanding our competence and implementation capabilities in the area of AI functionalities through our in-house AI infrastructure and the central RÖDL AI Hub. In addition, we are actively involved as a practical partner in a state-funded research project to take innovative approaches to improving the quality of AI results in auditing to the next level.

All our initiatives in the field of AI allow us as RÖDL to look optimistically into the future. We are able to actively shape the technological transformation in this area and thus make a significant contribution to the auditing of the future.