The future of AI for identity verification
Socure Founder and Chief Strategy Officer Sunil Madhu joined Arif Ahmed, SVP of emerging payments and technological innovation at U.S. Bank, and Frank Villavicencio, CPO for access and identity at ADP, onstage at this year’s Money20/20 to discuss AI’s role and future in identity verification and customer experience. The panel was moderated by Dan Rosen, founder and general partner of Commerce Ventures.
The panel started off discussing the age-old problem when it comes to identity verification: If there’s too much security-related friction, you lose business. But if the security’s too lenient, you’re a lot more likely to see false positives and bad actors. To that end, rules-based systems are no longer efficient—and can only protect you from things that they recognize as having happened in the past. At the speed at which fraudsters are evolving, only AI-based verification can work.
All the panelists also agreed on another surprising key point: As important as identity verification is, customer experience is king.
“Every business depends on customers,” Madhu says. “And customers are getting more and more impatient because every generation is mobile-first…. If you want to grow the top line, you have to care about acceptance, and fraud reduction is a cost you have to minimize.”
Ahmed and Villavicencio agreed, noting that it’s impossible to favor one over the other in this day and age when it comes to customer experience or security, and empathy for the user is integral to the solution. AI-based tools can do both, letting companies have their cake and eat it too.
As Madhu points out, in the past, identity verification has only been able to optimize for two of three sides of the triangle: acceptance, fraud rate, or manual review. AI allows you to optimize for all three, resulting in orders of magnitude in improvement.
The next steps for identity verification companies and those that rely on the process are fewer humans and more collaboration, the panelists noted.
Machines are simply better equipped to sort through the tons of data required these days, and the more that machines can learn and apply their learnings, the more accurate the results will be, eliminating human error and bias. In three to four years, they say, you can expect to see a lot fewer humans doing the work in this space.
As far as collaboration goes, technological consortia may be on the horizon as well. After all, if banks or other companies are already spending millions of dollars on KYC and identity verification, why can’t those verified users get the green light elsewhere as a result? Globally, this is already happening, and in the United States, banks are already thinking about ways to work together in this regard. The question of liability is still an issue, however.
But collaboration can have other benefits too.
“The whole notion of AI is collaborative filtering, collective intelligence,” Madhu says. “You want to protect yourself against the threats that you may face, but also against the ones that your peers in the industry may face as well.”
Therefore, the next generation of identity verification may very well see various companies’ experiences being blended to make AI smarter than ever and able to defend against ever-evolving new attacks.
Watch the entire panel discussion here.