A Montana startup today launched its flagship solution that assesses and predicts risk based on how people answer questions during an online application process. The technology, which relies on what the company’s founders call “prescriptive analytics,” is based on years of research into how people behave with an online user interface as they answer a series of questions—on a loan application, for example. In short: based on how users move their mouse or manipulate a touch screen and the answers they give, how likely is it they are being truthful on an online application? According to CEO Jack Alton and University of Arizona professor Joe Valacich, the company’s chief science officer and co-founder, Neuro-ID’s Confidence Score can do just that with 97 percent accuracy.
“When making underwriting or risk-based decisions, you now have a new attribute to look at: not just what they answer on a form, but how they’re going about it,” Alton said.
The company will target banks and businesses that offer loans, credit and insurance initially, but Alton agrees the potential in the payments industry, especially for payment providers onboarding merchant customers, is obvious.
The technology grew out of research by Valacich and colleague Jeff Jenkins, an IT professor at Brigham Young University, into understanding when online users were upset by monitoring mouse movement so someone might be able to provide a service intervention.
“The motor-nervous system—your hand movements—are tightly coupled to your brain,” explained Valacich. “As you increase stress, it manifests in changes in your hand. That led to years of research on tens of thousands of human subjects in different contexts. Our accuracy rates for detecting anomalies that reflect suspicious behaviors is near 100 percent.”
Companies can apply Neuro-ID’s Confidence Score to existing online applications and gain a better understanding of what questions an applicant might be lying on. Given a possible untruthful response, additional questions can be asked that can confirm the suspicious answer.