Skip to main content
search results
Sorry, but nothing matched your search terms. Sorry, but nothing matched your search terms. Sorry, but nothing matched your search terms.
Sorry, but we cannot handle your search query now. Please, try again later! Sorry, but we cannot handle your search query now. Please, try again later! Sorry, but we cannot handle your search query now. Please, try again later!
Search suggestions

Hard-working bees that get the job done quickly, reliably and without error  

Steffen graf, digitalisation expert

Steffen Graf is Cluster Lead Smart Automation at Commerzbank and Managing Director of Yellow Automation GmbH, a subsidiary of Commerzbank AG that specialises in the automation and operation of automated banking processes.  

In this interview, the digitalisation expert explains how artificial intelligence can help drive the automation of backoffice processes in the financial sector.  

How can AI support process automation in banking and financial services?

In back-office process automation, AI primarily plays the role of an enabler, a technology that prepares data for automation. This involves digitising and structuring information, such as a text-based letter or phone call. This is already possible with Large Language Models (LLM), with good results. The technology is also suitable for direct interaction with customers. An avatar can extract and structure information from a consultation, which in turn forms the basis for further processes. These enabler models are currently the main use case for AI in conjunction with Robotic Process Automation (RPA).
A digital representation of a human head made of a blue wireframe mesh dissolving into scattered pixels and geometric shapes, set against a dark background, symbolizing the concept of technology, artificial intelligence, or data disruption.

Can the use of AI be extended beyond enabler models?

Yes, for example, when the decision-making process in classic RPA applications is not clear. Until now, RPA bots have been rule-based. Depending on the situation, a decision is made on how to proceed. In the future, more powerful AI models could be used to make fuzzy decisions – decisions that do not fit neatly into a yes-no scheme but depend on other factors. But that is still some way off.  

“AI is making RPA technology faster and better.”

How reliable are the latest AI models?

We have been using text recognition software and related technologies for many years. This has been particularly useful for form-based text. However, with the recent technological leaps in the area of LLM, we are no longer limited to forms, but can extract information from a wide variety of documents or file types. This opens many more use cases and our pilot use cases are extremely promising.  

Will AI bring a quantum leap to today’s RPA technology?

We will be able to implement more and more use cases with AI as data processing gets faster and better, and as RPA technology advances. Programmers are already being assisted by AI with code snippets or code frameworks. While RPA is a specific use case, in the future AI could, for example, create an initial scripting framework for an RPA application and humans would only need to adjust, greatly increasing the efficiency of bot development.  

Is it possible for banks to achieve 100% automation?

This only makes sense in a few cases. Of course, many processes can be 100% automated. But the cost is too high. There will always be exceptions. If I implement a solution for every exception, the automation effort will be disproportionate to the efficiency gain. A level of automation of around 80% makes more sense. This allows you to automate standard cases and timeconsuming data entry. People are then only involved in more complex decisions and can focus on valueadding activities rather than routine tasks or data entry. With AI and RPA, it is possible to reduce the workload and improve the nature of the work. It is not without reason that Yellow Automation GmbH speaks of cobees, i.e. the bees of Commerzbank: hardworking helpers that actively support day-to-day banking and carry out many tedious tasks quickly, reliably and without error. 

With AI and RPA, it is possible to reduce the workload and improve the nature of the work.

This interview was taken from our recent report How AI is shaping the future of financial services.   

Download

Download whitepaper