Using AI to Manage Risk
For acquirers, managing risk has always been a moving target that requires paying attention to not only the transactions that come through, but also the payments landscape. Fraudsters are not deterred from trying everything they can to commit fraud, so the burden of protecting merchants and cardholder’s rests in risk management representatives’ laps. There have been many kinds of software in the past that has been used to help acquirers manage their risk, but what if there was a way to automate of the risk management process?
Chaitanya Vaidy sought to tackle this problem by creating DeepRisk, an artificial intelligence based end-to-end risk management platform for merchant acquirers. DeepRisk uses AI models to learn what risk is, how to identify it, and to recognize when a transaction is at risk of becoming a risky one. DeepRisk also automates the validation checks prior to boarding a merchant, by automatically running background checks, business verification checks, and more to determine if a merchant is a future risk or not. “Underwriters are risk analysists too”, says Chai, and by using DeepRisk, Underwriters take advantage of the AI software to run checks in less than 30 seconds versus spending 15+ minutes manually pulling those checks per merchant.
Let’s get into where DeepRisk really shines: risk management. DeepRisk can do the basics, like holding and releasing funds, but the real trick is the AI; DeepRisk learns more through use what looks like fraud and helps to identify it. Whereas previously, Risk Analysts had to dig through thousands of transactions looking for anomalies, DeepRisk can reduce that amount to only a few hundred or even less. The AI can ID those transactions that can seem suspicious but is able to determine that they are not. Anything that it cannot identify with 100% certainty is sent to a Risk Analyst to review. This is the important part, that the software is not direct replacement for human Risk Analysts. The AI model is dependent on learning from the behaviors of those Risk Analysts so it can continue to be an accurate model for future risk. This cooperative relationship upends of how we traditionally think of technology replacing people. DeepRisk is not a people replacement, it’s an enhancement to the experience. “One Risk Analyst can now do the job of 10, because all of the manual stuff, the easy stuff, has been automated away. Risk [analysts] can now really focus on the transactions that matter and keep overall risk at less that 1% or lower.”
When asked about the future of DeepRisk, Chai says his team plans to implement newer AI models that are more sophisticated at learning and continue to develop the technology beyond the applications of Risk. How we use AI in the future to change how we use our business is already proving to be the leading way that the payments industry will evolve.
To learn more about DeepRisk visit DeepRisk.ai.