AI use case: Efficient financial analysis with AI-supported chatbot at Possehl

This article has been translated for your convenience using machine translation. Reasonable efforts have been made to provide an accurate translation. The official text is the German version of this content.

As part of the Possehl Group, Possehl Digital is always on the lookout for digital issues in the industrial SME sector and tries to find answers and solutions that are suitable for SMEs. This was also the case when the holding company of the Possehl Group asked whether AI tools could be used to analyze the financial data of the more than 200 Group companies even more efficiently to reduce complexity in controlling.

In particular, corporate groups with a high degree of decentralization often face the challenge of analyzing financial information from different sources and systems. The distribution of this data makes it difficult to evaluate business performance quickly and comprehensively.

Financial Controlling at Possehl Holding was also confronted with these issues. The special feature of the Possehl Group is its division into a total of ten divisions and more than 200 companies worldwide.

All 200 companies prepare monthly and quarterly reports on financial and business development. This information is evaluated, compared and commented on at holding company level.

Due to the special structure of the group of companies, the data is available in different formats. This made manual analyses unavoidable and led to corresponding expenses.

To optimize this process, Possehl decided to work with DataSpark. DataSpark is a subsidiary of Possehl Digital and specializes in providing automation solutions for a wide range of requirements. The aim was to use an AI-supported solution to make financial analysis more efficient and facilitate access to relevant information.

Challenge

Complex analysis of heterogeneous data formats

Controlling at Possehl Holding was faced with various challenges:

• Complex data preparation: the preparation of reports and comments required manual, time-consuming analysis of various sources of information.
• Heterogeneous document formats: Different layouts and formats made it difficult to process the financial data quickly and in a standardized way.
• Short processing time: There is only a limited time window between receipt of the reports and the final analysis. 

Solution

Introducing an AI-supported chatbot

DataSpark's answer to these challenges was to implement an AI-supported chatbot that enables user-friendly and efficient interaction with the data sources. The solution is based on the GPT-4o language model and can extract, analyze, compare and summarize relevant information from various documents, such as quarterly reports and presentations.

In addition to text-based information, the AI is able to interpret diagrams and graphics from presentations, transforming visual content into meaningful insights. New quarterly reports are uploaded via an intuitive user interface so that the latest data is always available for analysis. 

Outcome

Less manual effort in financial controlling

The use of AI-supported chatbots led to a reduction in manual effort when searching for relevant information. The ability to search data in natural language and automatically analyze complex financial developments from different formats improved the efficiency of the evaluations.

We are constantly working to optimize and further develop our solutions so that we can continue to create added value within the Possehl Group and beyond.

By combining human expertise and machine intelligence, we have created an analysis tool that offers Financial Controlling new opportunities for decision-making – even with unstructured data sources.