Debt collection and digitalization: Innovation award for EOS Deutschland!

At an early stage, the FX team was joined by a test team, as Michael Robert reports: “We sought out internal staff who were willing to work with us on the FX development as pilot users.” These colleagues are making an important contribution to the practicability of the system, because they have already processed actual receivables cases with the MVP, the minimum viable product. Their experiences and feedback then flow directly into the ongoing development. “From the outset, everyone involved in FX was located in the same area. The communication channels are short and we all work together on an equal footing, now that’s what I call agility in practice,” says Robert. In the meantime, the five-strong test team numbers around 35 users. Apart from collection agents these now also include the first specialists working on more complex receivables cases. 

Only the agile development work between users, business, IT and data science experts is producing results that actually facilitate individualized yet efficient receivables management. Today, FX is capable of providing collection agents with specific recommendations on how to proceed. The AI experts call this determining the “best next action”, i.e. the most promising option. 
To this end, the collection software has been populated with a comprehensive list of measures from which it suggests the best option based on an analysis of the case. In the process, it draws on criteria like creditors’ instructions, legal limitation periods, available communication channels or previous reactions by the defaulting payers.  

“What we need to do now is to increase the granularity of these recommendations,” says Robert. “In addition, the intention is of course for FX to take over more and more work steps. The automation of processes continues to be an important issue in receivables processing.” Naturally it is also conceivable that insights from behavioral research could be incorporated into the AI in future. Because although people don’t always make decisions rationally their decisions are often predictable. So for example, the Software could determine the optimum level of installments that reduce the debt but at the same time do not overburden the defaulting payer. Michael Robert is convinced: “Above all we want to achieve a long-term debt reduction that actually works, and in this context AI can be very useful.” 

As a technology-driven financial investor and service provider operating in 26 countries worldwide, the EOS Group invested more than €25 million in innovative technologies in fiscal 2019/20. Around 500 employees worldwide are responsible for the ongoing development and implementation of digital processes. Moreover, EOS has been active on the market since 1974 and thus has extensive experience, along with access to a wealth of historical data that can be used to “train” an AI system. The company also has a corporate culture that allows it to adopt new approaches with an open mind towards new technologies. 

“All this combined creates a framework which gives us the freedom to develop an AI-based collection software system like FX,” says Michael Robert. Already it is clear that the working methods and technologies used for FX are also making their mark on other projects in the IT environment. “Even before FX, significant refinements were being made in other IT areas. These have now been given yet another massive push and more and more synergies are emerging between further developments in FX and the work of other teams.” 

Although initially only some new receivables were processed with the FX collection software, other receivables will now be migrated to FX in the course of the rollout, until in a few years time, FX will eventually have replaced the existing collection system entirely. And every step taken until then improves efficiency.