Artificial intelligence is having an increasingly significant impact on the way the largest financial institutions operate. Deutsche Bank has reported that, thanks to the use of AI tools, technology projects that until recently required up to two years to complete can now be delivered within three to six months. This assessment was presented by Denis Roux, Chief Information Officer at Deutsche Bank, during the Bank on Tech event in Bengaluru.
According to Roux, the effects are particularly evident in the area of outstanding technology work and internal processes. Tasks that previously took months to complete are now being finalised within a few weeks. The bank views AI primarily as a tool to boost team efficiency, rather than an end in itself.
At the same time, Deutsche Bank highlights the rising costs of using generative models. An increasing number of providers, including OpenAI and Anthropic, are adopting usage-based billing models, measured by the number of tokens processed. In response, the bank has introduced a system of limits for engineers. Additional resources are allocated only after the project has demonstrated specific business value.
This approach is part of a broader trend observed in the financial sector. Following an initial phase of experimentation, banks are increasingly focusing on measuring the return on investment in AI. The aim is not only to implement new solutions, but also to control the costs of infrastructure and computing power.
Deutsche Bank is currently developing tools for the automated analysis of financial data, document processing and the assessment of the impact of geopolitical and market events on its assets. The bank also utilises solutions that support analysts in creating reports and synthesising large datasets.
Despite its enthusiasm for new technologies, the institution remains cautious. Roux emphasises that not every task requires advanced generative models. In many cases, simpler algorithms or traditional IT solutions remain more effective.
The example of Deutsche Bank shows that, for large organisations, artificial intelligence is no longer an experiment. It is increasingly becoming a tool for streamlining day-to-day work, provided that its use remains economically viable.
