Spin Analytics automates credit risk modeling for banks

Meet Spin Analytics, a startup that wants to leverage artificial intelligence to automatically write credit risk modeling regulation reports. The company is participating in Startup Battlefield at TechCrunch Disrupt Berlin.

If you work for a big bank, you know how painful it can be to launch a new product. Every time you start selling a new asset, you need to comply with regulations around the world. It can take months and a lot of money to write detailed documents about your asset.

This isn’t like writing a school essay. You need to validate the model, stress test and make sure that everything is sound. “The idea is to automate this process. Today, this process takes 6 to 9 months,” co-founder and CEO Panos Skliamis told me before Disrupt.

Spin Analytics calls its platform RiskRobot. First, you need to get a clean data set. The startup helps you aggregate, merge and cleanse data before processing it. This process alone usually takes 4 to 6 weeks.

Second, RiskRobot makes sure you comply with regulations in Europe, the U.S. and all around the world — Basel III, CECL, you name it.

Finally, Spin Analytics writes the big report. Regulators want to make sure that it’s accurate. That’s why the report contains step-by-step instructions so you can reproduce the model later. Overall, you can expect to leverage Spin Analytics to write a report in less than two weeks.

Spin Analytics has been working on this product for three years and is now testing it with some big banks, such as BBVA and Crédit Agricole. If everything goes well, those banks could end up using Spin Analytics for more and more asset classes.

It’s an easy sell, as banks could end up saving a ton of money. Credit risk management currently costs $500,000 to $1 million per model. “We reduce that by 70 percent,” Skliamis said.

Now, banks need to assess the risk of using this credit risk modeling system. It sounds a bit convoluted, but it also sounds like a great business opportunity.