The Augmented Underwriter: can statistics, artificial intelligence, and quantum computing empower further Credit Insurers?

We are risk management professionals by nature and risk-averse by culture. Such a combination was meant to give little chance of success to innovative ideas and proposals. Yet, the rise of FinTech and InsurTech has recently forced us to revisit our value proposition. Following the step of denial and rejection well known by change management consultants - the industry has initiated all kinds of partnership approaches to minimize the damages caused by new players to the incumbent's value proposition. But have we really listened to the messages our customers and regulators were delivering?

Introducing credit insurance and science to each other

What matters is that trade and infrastructure finance gaps are growing, and like global warming, the more we talk about them, the bigger they grow. According to MIT, there lies a big problem, and big" I" innovation happens when big problems meet big solutions, i.e., brand-new technologies. Obviously, such a change does not occur overnight, and there is still value in small "i" innovation leading to incremental, short-term improvements that most executives are willing to finance. Big "I" requires allowing for failure, out-of-the-box thinking, and a genuine commitment to tackle current limitations.

Companies and regulators are not challenging the value proposition of Credit Insurance; they are challenging us for being an "umbrella that never opens." They are both right and wrong on this. Nevertheless, we do take up the challenge of doing more and better. They do have big problems, and recent technological developments in IT now open the doors for big solutions, so we must give them a try. Credit insurers are sitting on substantial data assets gained throughout decades of underwriting. Artificial Intelligence and Quantum computing scientists have made tremendous progress in recent years, powered by cloud computing commodities. How comes these two worlds have not met yet??

Tinubu's entrepreneurial culture and willingness to maintain an innovation profile have allowed me to staff our Innovation LAB with scientific profiles, which are rare in our industry. Investment banks hire mathematicians to work on options pricing, and specialty insurance is supported by actuaries, even if they can have difficulties understanding the products themselves. However, convincing an executive committee within our industry to approve a job description requiring an academic background in quantum physics is a different kettle of fish! Yet, Tinubu did it!?

Feedback on 12 months of research

The Tinubu's Lab Innovation white paper compiles a collection of experiments we have been running during the last 12 months based on our preliminary use case and the screening of promising new technologies. Not all of them offer promising perspectives, but the knowledge we have gained remains invaluable. As you will read throughout the following pages, the primary factor limiting the team's work is quality data and ambiguous validation criteria. In the end, the lowest grade given by credit rating companies corresponds to a back-tested 30 percent default rate, meaning that their default evaluation could be considered inaccurate in 70 percent of cases. Being provocative, it would mean that a machine learning model able to predict a company's default with a 40% accuracy rate would beat incumbent practices.

Unfortunately, it is not that simple. Research is not either correct or wrong; it is about learning, collaborating, and moving forward. We can all agree that we need more people with diverse backgrounds to join our industry; scientists are one kind of them. We have considerable assets to attract them with our data, and we could certainly get more results by acting together rather than separately.

Again, research is a long-term endeavor, and on top of the sprint concept, which is the backbone of modern software development practices, the leading software in credit insurance is allowing itself to invest in research. The use cases we have been working on need supporters to bring them forward, so feel free to pick one (or a few) of them and help us get closer to production-ready offerings. We will welcome any player willing to join us on this journey.

Take a look at the white paper and enjoy a good reading time.?

By Thomas FROSSARD, Chief Innovation Officer