Step by step to the future of credit management

In the last issue of the KPMG Corporate Treasury Newsletter from April 2025, the question of when credit insurance is worthwhile in operational credit management and how a risk portfolio can be managed was analysed in detail.

In this article, we now present technical solutions for credit management and analyse how the use of artificial intelligence (AI) can change processes in this area in the future.

The aim of digital credit management is to create an information base that enables a risk to be assessed at any time and decisions to be made on this basis. In order to enable more decisions to be made in the same amount of time compared to manual processing, it is proven that data is aggregated, analysed and visualised in a software solution - ideally automatically, so that the required information is displayed without being requested as soon as changes occur and the process of making enquiries is no longer necessary. Technical solutions show how debtors are distributed across rating classes and how important key figures develop over time. They also help with automated processes that can be customised and simplify your processes. The combination of market experience and targeted data analysis forms the basis for successful credit management software. With regard to the risk profile of a debtor, a differentiated view is also required, for example to distinguish new customers from existing customers in the evaluation, as different detailed information bases can be included in the evaluation. As part of the development of suitable information bases for individual sectors, data suppliers are a starting point, providing corresponding data via their interfaces, supported by credit management software. This data can in turn be weighted in the company's own context in order to assign corresponding risk scores to its own debtors. In this respect, modern solutions offer the advantage of customised mapping of individual credit policies and flexible adaptability. To summarise, the use of a software solution pays off in terms of the company's results, as you can automatically implement the selectivity of your risk portfolio in accordance with your individual credit policy.

So what role does AI play in the context of such credit management software? A classic operational solution implements the individual requirements of a company's credit policy and automatically ensures the quality and timeliness of information so that well-founded decisions can be made, measures taken and errors avoided. As the amount of data used today for decision-making is constantly growing, AI is an ideal solution for credit management, allowing classically manually evaluated rules to be learnt by a model as a result of training on the available data. This offers the advantage that rules and patterns that were previously mapped in individual credit policies through expertise can now be created on the basis of the company's own objective data and the effort required to identify them can be automated. This places a strong focus on data quality requirements, which can generally have a positive impact on information management in companies in addition to the development of AI strategies in credit management¹. In addition to the classic as-is analysis, AI also offers the potential to make statements about the future and to automate manual routine activities that previously seemed too complex using generative AI. Depending on the process in credit management software, it is important to consider which task the AI supports. If it is an evaluation of structured data, machine learning is suitable for training a model, while language-related tasks, such as summarising texts or extracting information from a text, can be mapped using language models (e.g. Mistral Large or GPT-4o). A concrete example of its use is risk analysis based on texts, such as news articles, social media posts and other publicly accessible company publications, which previously had to be analysed manually. This opens up ways of not only incorporating new information into the decision-making process, but also being able to record it directly and automatically.

As is common in the software industry, the success of credit management software generally depends on its user-friendliness for the end user. In the transformation of credit management based on the use of AI within software, the focus is not only on automated decision support, but also on traceability and transparency. The aim here is to utilise procedures and explanatory models in such a way that they are comprehensible to end users. The EU AI Act suggests that employee training should ensure that both the potential and the limitations of AI-based outputs are recognised. Humans will therefore clearly continue to play a central role in credit management, but their activities themselves will change. While the time-consuming procurement of information, research and checking of information was previously part of the job, AI offers the potential to process precisely these tasks as individually as a credit decision-maker wishes in a fraction of the time previously required. Accordingly, a credit decision-maker can use AI to create the required information base so that the decisions to be made remain comprehensible and can be made much more quickly.

In principle, it can be said that the software basis in credit management does not need to undergo any comprehensive change, as it already ensures that a process-orientated data strategy can be pursued, which forms the cornerstone for the successful use of AI. However, the extent to which software will be allowed to make automated decisions in the future will continue to depend on the assigned competences and regulatory restrictions. Experience with and understanding of AI will therefore be drivers for trust and therefore also for the design of its use to transform traditional processes in credit management. From today's perspective, it can therefore be assumed that the future will also be shaped step by step in software - through the continuous further development of processes and the targeted, gradual use of AI as a strategic player.

Source: KPMG Corporate Treasury News, Issue 154, May 2025

Guest author:

Dr Tobias Vinhoven, Product Owner AI,

Prof. Schumann GmbH

__________________________________________________________________________________________________________________

1 Source: https://klardenker.kpmg.de/der-schluessel-fuer-geschaeftserfolg-hochwertige-daten/ (accessed on 11/04/2025)

The views and opinions expressed in guest contributions are those of the author and do not necessarily reflect the views and opinions of KPMG AG Wirtschaftsprüfungsgesellschaft, a stock corporation under German law.