Group
It’s clear that artificial intelligence (AI) is gaining more and more ground. But how will this technology influence existing industries? And what will be the consequences for the activities of credit managers? “AI will accelerate many activities, including credit management, and it will re-shape industries. The dominant AI activity has been towards increasing productivity and quality, using technology to perform roles like collecting and analysing data, predictive analyses, because AI is far better at recognising patterns than humans and it processes enormous volumes of data in near real time. Generative AI has taken AI capabilities to a new level with its ability to generate text and images in a way that mimics humans. These technologies will help credit managers do their jobs better, but we don’t see it fundamentally changing the role of credit managers – yet,” says Stan Chang, Director Group Buyer Underwriting at Atradius.
Chang: “AI has been an integral part of our toolkit since 2016. We integrate AI and Machine Learning (ML) into our ecosystems to achieve more speed, quality and efficiency in portfolio management and credit decision-making. These activities include automatic retrieval and analysis of information from numerous sources, entity matching, machine reading of financial statements, and we use neural networks and fuzzy logic to optimize credit decisions.”
The results are impressive, it turns out. “We have achieved significant efficiency improvements, productivity increases, improved customer service and claims savings. This is especially important given the enormous size of our risk exposures and the millions of credit limits we process,” Chang says.
The implementation of AI brings challenges similar to the development of technology tools. Chang emphasizes the importance of knowledge: “The first step is to learn. While technical knowledge is not needed, users should try to understand what AI can and cannot do, where it’s likely to produce results and benefits, and identify use cases. This is still very new and exciting. The great news is that (Generative) AI has become widely accessible to the mass market at consumer and business level.”
Chang advises starting small: “An effective way to overcome challenges is to apply AI tools to familiar processes where you can assess performance and output. Start small and don’t try to change the world right away!”
He also warns about the complexity of the AI market: “Choose your partners and vendors carefully. There is an enormous community of providers, consultants, fintechs and experts – AI is big business and everyone is riding that wave. While AI tools are becoming increasingly generic, we’ve found that customization is needed, for example in specific use cases, processes and desired user experience.”
Chang is convinced of the continuing relevance of credit insurers: “The role of credit insurance is to compensate customers in case of non-payment for goods or services provided. The risk of default doesn’t decrease with the use of AI, even if AI can help us identify and understand those risks more comprehensively or accurately.”
He emphasizes that it ultimately comes down to risk appetite and business needs: “Being aware of risks, for example in transactions with a buyer in the Middle East, doesn’t mean that payment is guaranteed, and it doesn’t mean you decide not to do business there, especially if the profits are sufficiently attractive. Credit insurance gives your company peace of mind in case you (and your insurer) want to take that risk.”
Chang also points to the broader economic importance of credit insurance: “Keep in mind that credit insurance protects accounts receivables. And accounts receivables are often the largest asset for many companies, against which many financiers lend money. So yes, we believe we will continue to add great value in the future.”
Chang sees possibilities for lower premiums, but places this in a broader context: “Potentially yes, and that would be as much due to the maturity of the industry, including the effective use of technology to drive efficiencies. But premiums are driven by many factors such as services, loss experience and risk outlook, market forces (demand/supply/competition/substitute products), and the cost of capital, which – especially in the current environment – may lead to an upward movement as well.”
Chang warns against hasty implementation of AI due to the hype: “AI is becoming widely accessible and affordable for companies, but it’s not about AI or GenAI itself. There are many use cases where technology is deployed without it being commercially justified. And while AI will likely form an important part of our future, we shouldn’t ‘implement AI’ simply because of the hype.”
Instead, he advises: “Look for ways to use technology to modernize your processes and services, and that improve your efficiency and competitiveness, such as digitization of documents, data management, streamlining workflows and connectivity tools. By the way, many of these elements also enable the use of AI. Start small, focus your scope, seek external help – it usually goes much faster if you do it well – and achieve early successes to build confidence.”
Concurrently there is an abundance of AI and GenAI information, articles, blogs, webinars, conferences, forums and communities who thrive on sharing knowledge and experiences. While propaganda abounds, it is nevertheless an effective and fun way to learn and to get ideas.
Chang sees a broad awareness of the benefits of technology, but notes that the motivation to experiment and adopt is often limited: “I think there’s widespread awareness about the benefits of technology. But the motivation to experiment and adopt is often hindered by the lack of a learning or change culture, or because people rely on practices they’re used to. In short: habits, resistance to change, and maybe even fear of technology stand in the way of innovation. Compliance with regulation and data privacy are also paramount in the consideration of using AI tools”.
Chang emphasizes that AI is a means, not an end in itself: “All tooling should start with the question ‘What problem am I trying to solve?’ rather than ‘How can I use AI?'” Nowadays however, with easy accessibility to and the possibilities of GenAI, it’s fun and productive to think about what can I use GenAI for!
I would encourage companies, small and large to promote innovative thinking and energy into their teams. We have found that it’s more effective to stimulate innovation through ideas and activities led by the business, with support from the IT department and funding from stakeholders.
“At Atradius, AI has powerful benefits because of the high volume and complexity of the risk exposures and workflows we manage.” He provides concrete examples: “GenAI, for instance, is able to distill large volumes of data from diverse sources into an easy-to-read summary of key considerations for credit assessment. We also use AI to make better credit decisions because of the enormous amount of internal and external data we possess, and the models we build.”
Chang sees AI primarily as a supporting instrument: “The key role of credit managers is to reduce DSOs, improve cash flow and minimize bad debts while giving the company room to sell more. AI can provide tools to assess creditworthiness, deliver portfolio insights, investigate portfolios (for example to identify emerging risks) and improve dunning activities or automatically trigger credit stops, especially in large companies.”
However, he emphasizes that these tools improve rather than fundamentally change the way credit managers do their job: “These tools improve the way credit managers do their job by bringing more quality, speed and efficiency, but I don’t see that they will ‘drastically change the role of credit managers in the future’. At least not for now, in my opinion.
Source: Creditexpo.nl
