By Dale S. Laszig
The richly layered payments ecosystem is mostly invisible to the consumers and merchants who use it every day. In contrast, simplifying payments has been a complicated endeavor for technology companies, financial institutions and service providers. This three-part series has chronicled the journey from payment acceptance to commerce enablement, recognizing important milestones along the way.
Part 1 recalled the initial leap into electronic transactions. Part 2 followed subsequent developments in mobile payments and ecommerce. Part 3 goes behind the scenes to speak with innovative designers and developers who are nurturing tomorrow's technologies.
Electronic transactions have played a pivotal role in the global migration from analog to digital technologies, experts say. As payment technologies evolve, previously defining characteristics, such as authorization and settlement, can become apps running in the background. And as they master language and cognitive skills, solutions move from basic tasks to interactive communications.
"Technologies that simplify payments offer merchants a single point of command-and-control where they can oversee transaction processes," said Vlad Branin, Vice President, Professional Services at Zooz Inc., a global technology company. "This gives them the ability to act in real time to rescue and optimize transactions. They can establish rules to govern each payment's route without manual intervention, reducing declines, downtime, and false positives, and ultimately increasing revenues and acceptance rates globally."
As payment technologies mature, they are never far from an equally robust, evolving threat landscape. Alex Vaystikh, Chief Technology Officer of SecBI Ltd., urges business owners to think beyond traditional security methods. "The old paradigm of perimeter, gateway, endpoint and signature-based detection is not relevant anymore," he stated. "New, more advanced techniques are now required to stop these evolving threats."
Vaystikh said purpose-built algorithms are replacing obsolete security schemes. The sandbox, a former gold standard in malware detection, provided a contained environment for observing suspicious files. Malware has evolved to the point where it knows when it's being executed in a sandbox environment, so it remains dormant for a long time, he noted. "Today's sophisticated solutions examine activity throughout an entire network to identify suspicious activities instead of relying on a single set of behavioral attributes," he said.
Professor Ronald Coifman, Yale University faculty member and renowned mathematician, received the National Medal of Science from President Clinton and is co-founder of ThetaRay, a data analytics and cyber-security company. "Human beings created machines in their own image, with eyes and arms," he said. "Artificial intelligence, used by leading financial institutions, identifies unexpected things by mimicking nature."
Mark Gazit, ThetaRay Chief Executive Officer, added, "Artificial intelligence can be a misleading term, because computers do exactly what humans do. In the early days, without much data or computing power, organizations used small algorithms to identify ever-present threats."
Recalling early development efforts, Coifman said, "You build a human model with uniform information about arms, legs and weight. But the best model of human beings might make a basketball player look like an alien, or a small animal look like a human baby."
The engineers decided to forego measurements and focus on proportions. This way, a basketball player would still register as a human baby, and a tiger would look like a cat. In a similar way, advanced fraud detection schemes are taught to differentiate between normal and unusual activity, without relying on static information or rules engines.
Machines and humans run algorithms to connect events with observations, but machines are faster and more efficient at processing situational awareness. "All the algorithms we have running are built of empirical linkages, based on experience, that provide useful information," Coifman said. "As we acquire data, our situational and quantitative awareness becomes more and more precise."
He said that it's not computing in the usual sense; it's about sensing by making connections. Every living organism has its own collection of sensors and linkages, and learns by making connections and measuring the strength of those connections to reach more precise conclusions.
Advanced threat detection begins with a blank canvas and builds observational models to differentiate between normal and suspicious behavior. These models don't operate on the same time scale as humans; they compute with dizzying speed and change all the time.
"By connecting all the dots all the time across multiple dimensions, artificial intelligence can process more information than human beings can capture in real time," Gazit said. "Despite all the progress in the scientific community, many security analysts still work semi-manually."
Using the analogy of ancient cartography versus the global positioning system (GPS) to compare old and new security methods, Coifman said the GPS uses precise coordinates to portray reality, while traditional maps that quantitatively measure distances on paper do not account for the physical relief of sea, islands and mountains.
Organizations learn in much the same way humans and machines do, said Mark Cerminaro, Chief Revenue Officer at RapidAdvance, an online financial services company. He said initial steps typically involve assessing and organizing data to improve efficiencies.
"Ideally, the objective is to structure the environment at a macro level to see what is operating efficiently, and where we need to dig in more to create continuity," he said. "Knowledge is empowering for an organization. We encourage our team to assess operations, interpret trends and share their opinions about what the data is telling them."
Cerminaro observed that financial institutions are stepping outside of their comfort zones and entering the world of technology. Their collaborations with fintechs are helping them leverage the traditional banking infrastructure and improve risk management practices.
The world has evolved; every partner has different infrastructure, resources and markets, Cerminaro said. When you look at the processing world, ISOs and independent software vendors are adding alternative lending solutions to their suite of solutions. We want to build true relationships with ISOs, so we brought in Jim Fink, who understands that business, he added.
Jim Fink, Director of Strategic Partnerships at RapidAdvance, concurred, stating, "In the history of our space, it's always been clicks. Today's processors are getting away from that model, which has become a race to the bottom. They understand this is no longer a transactional business and see the value of being a full-service provider to their clients."
The software realm is adapting to emerging technologies and faster development cycles. For years, a software development kit (SDK) was simply a set of tools backed by email support. Today's developers expect real-time support services and robust application programming interface libraries that support seamless integration. "Twenty years ago, Steve Jobs said the computer needs to evolve from an object to an appliance," said Mike Gardner, CEO at Agreement Express, a software-as-a-service company.
"To paraphrase Jobs, when technologies are innate skills, they become apps running in the background. This is true for all technologies, not just payments." Gardner has seen agile (iterative) development replace waterfall (sequential, non-iterative) development over the past 16 years. "Waterfall development was a World War II construct, and agile was post-Vietnam, from the Air Force," he said. "You can't take a rigid approach to combat; you have to adapt to circumstance."
Among its setbacks, waterfall development was cumbersome and slow. "Being overtime, over budget, over engineered, the first generation was awful," Gardner said. "Imagine telling a consumer, 'If you download this app on your phone, within six to nine months you'll be able to use it, and by the way, you'll have to spend $150,000 and when I bring it live, it may not include all the things you wanted.'"
Gardner went on to say that payment technology must be conducive to agile implementation. Agreement Express spent five years re-engineering its platform and brought it live in November 2016. The platform learns, adapts and makes data-driven decisions, he noted. Over time, it provides business owners with better and better information, he added.
All enterprise technology needs to get to this point, Gardner said. When I'm traveling, I may need a translation app to read a menu. When we say, "It will take six to nine months to activate this," we're creating tomorrow's legacy system. "It's a fascinating world because technology must be seamless at the enterprise level," he said. "We're at the beginning of what we're going to see in the Internet of Things (IoT). And the computer is still not an appliance, not yet. If you think about the IoT, the appliances have become computers."
J. Scott Sanchez, Innovation Leader at First Data Corp., said that in addition to providing a secure, convenient experience, payments have moved from transactions to interactions. Conceptually, it's about facilitating the right interactions between people while money moves in the background. "We're in a good position to innovate, but I wish we could just skip to seamless," he said. "Call it ease of use or intuition; the idea is using something without noticing that you're transacting."
Five years ago, user-centered design thinking was new to First Data, Sanchez said. Today it's implemented across the organization as product teams work with banks, shops, ISOs and ISVs. We have a saying: quick it or quit it; if you're not quick, they will leave you quickly, he said.
Sanchez, who formerly ran Intuit's Innovation Lab and served as Vice President, Global Innovation Strategy at Visa, said that ultimately the objective is to give people things they could never imagine, then never imagine going without. Sanchez also stressed the need for repeatability. "Getting paid once is not the idea; it's getting consumers to come back."
Following are three complementary criteria of the design thinking process:
Sanchez said product managers holistically manage all three criteria.
Designing for the real world requires empathy and pragmatism, Sanchez noted. Researchers visit people in their homes and conduct research in the field to solve problems the right way, but also to solve the right problems.
"When thinking about building a mobile wallet for gas stations, my chief product officer and I went into quick service restaurant [QSR] and petroleum verticals to study people," Sanchez said. "It turned out the payment itself wasn't broken from a consumer perspective. Payment cards worked well. But we found other problems in the drive-through."
After observing the high number of queues and extended wait times at gas station QSRs, First Data designed its Order Ahead app. People could order on the app, pick up their food, and payment and loyalty just happened as part of the experience.
When a customer in a sporting goods store says, "I think I want this kayak," the distance between the checkout queue and that impulse decision is payments' critical last mile, said Allan Lacoste, Director of ISO at Total Merchant Services Inc. A tablet-carrying sales associate can act on and reinforce that impulse.
"What doesn't work is a solution that requires exclusivity and unfairly restricts our sales partners from making a living," Lacoste said. "Proprietary systems are out of place in a world of connected cars, dishwashers that order detergent and Amazon Go! markets where you can pick up your groceries and leave without taking anything out of your pocket."
He added that the IoT has created a whole new runway, wide open for growth. "It will be interesting to see how far we can take this and how the payments landscape will look in 15 years," he said.
Notice to readers: These are archived articles. Contact names or information may be out of date. We regret any inconvenience.Prev Next