Treasury is too complex to build on your own
DerTreasurer asks whether artificial intelligence will render treasury management systems obsolete. Thomas Dohmen and Michael Völkl respond: No - something else is the key factor.
Contents
At a user conference, a DAX-listed company demonstrated how its treasury department used Python and AI to build its own solution for reclassifying hedges. This eliminated about 200 manual steps per month, reducing the process from two to three days to just a few hours. For the trade publication DerTreasurer, this raised a fundamental question: Will treasury departments simply build their own software in the future? In Jakob Eich’s article, our CEO Thomas Dohmen and Michael Völkl – who is responsible for treasury offerings at BDF EXPERTS – share their perspectives on this topic.
The competitive advantage lies in integration, not in the tool
Treasury systems thrive not on their user interfaces, but on what lies behind them: the integration with sales, materials management, finance, and payment processing. The TMS draws information from these processes, which it uses to generate recommendations. According to Thomas Dohmen, it is precisely this integration that represents the true value – and the reason why AI will not replace these systems or their providers in the long run. However, this requires a robust, reliable database within the company itself. Dohmen calls this a “Clean Data Base.”
“In my view, it’s more of an add-on than a game-changer.” Thomas Dohmen, CEO of BDF EXPERTS, in DerTreasurer
AI can support, simulate, and prepare decisions. It doesn’t go a step beyond that.
The hurdle is experiential knowledge
Michael Völkl takes a different approach: Treasury relies on know-how and experience, and every industry operates differently. Anyone who wants to build their own AI-powered system would first have to make this knowledge machine-readable.
Dohmen sees this as the real challenge – and it lies with people, not technology: Treasury professionals would need to be able to articulate what their expertise consists of and how they make decisions in specific situations. In practice, this rarely happens. Only on this basis could IT build a specific solution. Methodologically, therefore, the path involves translating the “best-of-breed” approach into software.
Accountability: Whoever Makes the Decision Must Be Able to Trace It
In treasury, it must always be clear who is responsible for a payment. If this responsibility remains with a person, they must be able to review and trace the results – and this is where many common AI models still fall short today, because they do not disclose their reasoning transparently enough. Michael Völkl draws a clear line here: support for decision-making, yes; AI making its own decisions in treasury, no.
This aligns with the principle by which we design AI functions in our own solutions: The AI makes suggestions; humans make the decisions. Every suggestion remains traceable, every transaction is documented, and every approval is tied to a specific role.
Waiting it out is the most expensive option
There is currently a great deal of uncertainty in the market. Dohmen observes that companies are postponing the implementation of a new TMS or switching providers because they want to see first where AI is headed – a trend that is also noticeable with SAP TMS, where this is causing a backlog.
This strategy rarely pays off. After all, what every future AI feature will require can already be established today: integrated, up-to-date, and verifiable financial data. That’s exactly where our solutions come in. With the Cash Position Cockpit (CPC), treasurers can view their liquidity position in real time and fully integrated with SAP; with the Liquidity Planning Cockpit (LPC), they can continue planning based on the same data – without separate lists or duplicate data.
Conclusion
After reading this article, we can confidently answer the question posed in the title: The treasury management system isn’t going away. The tasks are too complex, and the responsibilities too critical. What is changing is the way treasurers work with the system – and the demands placed on the underlying data. Those who get this foundation in order now will be able to leverage AI functions as soon as they are reliable. Those who wait will simply have to wait longer.
Read the full article at DerTreasurer → https://www.dertreasurer.de/news/treasury-management-systeme-tms/laeutet-ki-das-ende-der-tms-ein-53708/
How robust is your treasury data foundation? Talk to us.
The text and graphics were created with AI support and reviewed by BDF EXPERTS.
Related articles
Treasury is too complex to build on your own
🎬 25 Years of BDF EXPERTS: What Sets Us Up for the Future (16/25)
🎬 25 Years of BDF EXPERTS: Ready for the Next 25 Years (15/25)
Finding
the rightsolution together.
We’re happy to answer your questions and help you find the right services to meet your needs.
