As eCommerce has exploded, payments have become more than just transactions, and more than just the last point of interaction between merchants and consumers. Payments have emerged as a force multiplier to drive new businesses and to help firms monetize data, as new commerce ecosystems are constantly being forged and expanded.
For the enterprise, though, the digital-first economy has increased the number of card-not-present transactions. That development has presented a challenge in ensuring that good customers and transactions are getting through and that conversion rates are as high as they can be, while the bad actors are being turned away.
That challenge is important to the continued development of the digital-first economy, and was a key topic in a conversation between Karen Webster and Neuro-ID CEO Jack Alton for PYMNTS’ Connected Economy series. From Alton’s perspective, actively monitoring and analyzing digital habits – via taps, types and swipes – can help secure connected ecosystems and help merchants fine-tune their offers and products in real time. At a high level, the role of data and artificial intelligence (AI) in making those ecosystems smart, safe and monetized can be viewed through the prism of a single question: How are we doing so far?
Alton sees a need for improvement. As he told Webster, firms that have migrated a large percentage of their businesses online have had to keep their “fraud guards” up. As payments have moved up and to the right — exponentially so — there have been plenty of opportunities to improve the consumer experience and root out fraud more effectively than what has been done to date.
As retailers are taking consumers into new adjacencies and finding opportunities to engage with them in new contexts, said Alton: “Everybody’s looking for new sources to try to replicate these human interactions that we used to have” in brick-and-mortar settings.
Replicating those interactions depends on one thing: data. Simply put, third-party providers or credit bureaus look at historical data and then try to predict the future. The process determines whether customers are low-risk or high-risk — or whether they are who they say they are. In an in-store setting, verbal and visual cues would help retailers make decisions about risk in the literal blink of an eye, enabling them to forge relationships with their customers on a daily and long-term basis.
“And when you talk about the ‘secret sauce’ to get us back to that point, to where these collisions [failed sales conversions] aren’t occurring, where we’re not saying no to our good customers and inadvertently saying yes to our high-risk customers … it really comes down to tapping into a new source of data,” said Alton.
Against that backdrop, AI- and machine learning (ML)-driven models need the taps and swipes to detect the behavioral interactions that consumers have with brands every day. These interactions can tell the brand something new every day — whether the consumer is enjoying the experience, are frustrated or might be trying to mask their identities.
Drilling down into those interactions, said Alton, firms can see where consumers are abandoning the engagement, at the point of onboarding or beyond, before committing to transact with that enterprise. It may be one document that needs to be uploaded, or a step-up during the login that proves to be one step too far for the consumer.
That type of analysis takes the friction away from fraud prevention teams, said Alton, so they can systematically go about reducing friction for their genuine customers. They can also route the appropriate level of friction for customers who may not be interacting with their data competently (and thus throwing up red flags to the companies). Beyond merely battling fraud, said Alton, there exists the opportunity to “nudge” the would-be customer toward completing their online journey and improving a firm’s drop-off rate.
“The challenge has been figuring out when to do that and who to do that with,” said Alton, adding that “we have been looking at the behaviors of genuine customers, finding out where there’s frustration, hesitation or indecision, or where people are getting lost or confused.”
That’s the appropriate time to nudge them forward, either with a chat or an in-person intervention, said Alton. He gave an example where a Neuro-ID customer, an online lender, was asking a consumer what their annual income was — but the borrower was asking the lender to look at their hourly income.
“There was a lot of hesitation, there was a lot of pausing,” he said of that online interaction. “There was a lot of manipulation of the answers, so that without really understanding the question and the context, it may have looked like fraudulent behavior. In fact, it was just a person who was confused, and trying to do the math on extrapolating their hourly wage to an annual income.” But adjusting for such activities — and reducing the friction in that data field — means the drop-off rate was reduced by about 40 percent, said Alton.
Neuro-ID also offers its clients a Friction Index Dashboard that can help them gauge the seamlessness of their customer journeys and smooth out any rough spots, Alton noted.
“We have customers that are leveraging the technologies that have experienced a doubling conversion of points in their customer journey that used to have 30 or 40 percent abandonment,” he said. “You have one chance to make a great first impression. That’s not just in person — that’s true digitally, too.”