Artificial Intelligence (AI) is no longer the future, but the present. AI is the technology that powers the verbal commands you give Siri. It operates driverless cars and can beat you in online games. It may sound intrusive, or even creepy, but companies are investing in AI to learn your behaviors – and monetize those behaviors.
So what exactly is AI? Artificial Intelligence is the simulation of human intelligence processes by machines – especially computer systems – to make decisions like a person. These processes include learning, reasoning and self-correction. Particular applications of AI include expert systems, speech recognition and machine vision.
AI is commonly used in smartphones, reservation systems, medical diagnosis and, of course, finance. It makes possible greater efficiencies, faster processing and almost complete accuracy, and is affecting nearly every industry, including the payments ecosystem.
The next few years will see an acceleration in developments that will alter how payment companies large and small operate their businesses. This is especially true in the high-risk merchant processing side of the industry, where the potential for fraud and high levels of chargebacks is rampant. Those high-risk merchants who take advantage of the growing suite of AI driven tools to reduce risk while providing a better customer experience are the ones who will thrive.
How AI is Powering Payments
It seems only natural that AI would move into payments. It has the potential to streamline the payment chain, reduce fraud, improve customer service and optimize risk assessment—all reasons venture financing is flowing into AI.
According to a CB Insights annual AI funding report, funding in AI start-ups increased from just under $600 million in 2012 to $5.021 billion in 2016. And in 2016 alone, over 200 companies raised $1.5 billion of equity funding in the first half of the year. This tremendous growth is spurred by the promise for radical change that AI brings to most industries.
There are many types of AI, but for payments – particularly the high-risk merchant processing industry – the most important types include Machine Learning, Natural Language Processing, Verbal and Vision Recognition, and AI bots. PointofSale.com’s recent article on AI provides a brief description for each and why they are important for payments in fighting fraud.
Machine Learning: a process where a software model is constructed and vast volumes of data are compared against it. Learning algorithms enable the model to evolve in response to changes in the incoming data. Machine Learning is crucial to the high-risk processing industry’s ability to detect and respond to fraud.
Financial services companies develop Machine Learning models with rules that outline typical fraudulent activity. The software can review data as it comes in and flag patterns that match the fraud model—indicating the transactions should be investigated. As more and more transactions and patterns are examined, the model can develop itself based on learning algorithms. Machine Learning models can also help high-risk payment companies optimize product offerings, detect money laundering schemes, and perform online risk management.
Natural Language Processing (NLP): allows a machine to understand spoken commands or questions. Among the most well-known applications using NLP are Apple’s Siri and Google Translate. In the processing space, NLP is being used to reduce customer service costs by taking customer questions without human interaction. As well, NLP is being used by Apple to allow iPhone users to initiate payments using voice commands. This reduces customer friction and makes the sales process easier.
Visual Recognition (VR): allows the software to identify objects. In the payments industry, this is particularly applicable to fingerprint or facial recognition as second level authentication methods. The technology enables customers to easily make mobile payments and provides security from fraud at the point of sale. Apple Pay uses this type of AI to secure POS payments by requiring fingerprint ID for each iPhone or iPad transaction. In the high-risk processing world, VR is important because it’s authentication capabilities reduce chargeback rates for companies targeted by fraudsters.
AI Bots: apps that present the appearance of being human, while being completely automated. They combine NLP and Machine Learning code to present effective interfaces that can handle thousands of interactions at the same time, rather than the one at a time limitations of human customer service reps. AI bots are used in chat software, messaging apps, and virtual assistants, and allow customers to make direct payments without leaving their preferred social media platform, or via text messaging.
In the high-risk space, AI bots can replace human service reps in handling customer questions and complaints. They speed up the service process and create significant savings.
How Merchants are Using AI
A perfect example of a nimble company using AI to grow their business is Walmart. With 11,700 stores and 140 million weekly shoppers, Walmart has used machine learning to build a digital relationship with in-store customers, becoming the second largest online retailer.
The big-box retailer has deployed associate delivery programs, which use machine learning to route deliveries more efficiently. Customers can also shop in-store or can order online and pick up at the store, avoiding checkout lines.
Laurent Desegur, Vice President of Customer Experience Engineering at WalmartLabs, states that building that relationship depends on a few factors: offering low prices, managing risk, and not only helping shoppers find the products they are looking for, but making that process as convenient as possible.
“We’re essentially creating a bridge where we are enhancing the shopping experience through machine learning,” Desegur said. “This is what we call the digital relationship. We want to make sure there is a seamless experience between what customers do online and what they do in our stores.”
Additionally, Machine Learning has helped Walmart to create “the store of the future,” drawing insights from customer data to improve personalization and providing recommendations that anticipate what shoppers are looking to purchase.
Their strategy seems to be working, as Walmart’s ecommerce revenue has risen 63% year over year.
As AI Evolves, so Does Cybercrime
The merchant processing industry is evolving. Adoption of AI solutions for improving the customer experience and reducing fraud and chargebacks is improving efficiency and profit margins. Those merchants and payments companies who are quick to adopt the latest AI solutions are more likely to succeed in the new, automated business environment. Those high-risk businesses who ignore the changes reshaping the industry run the risk of being left behind by their more nimble competitors. In an environment where competition is increasing and margins are tightening AI provides high-risk companies new opportunity to evolve and grow.
It is important to all industries for the growth of AI to continue. But an interesting take on AI in a recent Pymnts.com article does provide a dose of reality when it comes to the speed of the evolution of cybercrime – and some security experts claim that AI alone will not be sufficient in protecting against data breaches.
G DATA researchers recently found that last year, a new malware specimen surfaced every 4.6 seconds. In the first quarter of 2017, it reduced to every 4.2 seconds, meaning millions and millions of new malware versions surfaced every year. Data from the IT-Security Institute found 127.5 million malware samples last year, and while there is evidence that number may decline for 2017, researchers warned that these cyberattacks are becoming more sophisticated.
With this volume of cybercrime, and because with machine learning you still need human involvement to detect the algorithms when a new type of attack emerges, machine learning can’t fight malware on its own.
Guy Caspi, CEO of cybersecurity company Deep Instinct, agrees that machine learning is not enough in an age of unprecedented evolution and volume of cybercrime.
“In machine learning, you still need pre-processing of someone who knows very well what you want to implement. For example, if you’re looking at facial recognition, you need someone able to identify, in every face, the specific features that are important to differentiate this face from any other face. This process is still done manually.”
Automation of AI will continue to be an important step to reducing fraud, but with malware being the usual culprit of data breaches, it will be crucial for companies to adopt security solutions that prevent malware from infiltrating a network or system. That solution, as the PCI Security Standards Council states, is PCI-validated Point-to-Point Encryption (P2PE).
Bluefin’s PCI P2PE solution encrypts cardholder data at the Point of Interaction (POI), preventing clear-text cardholder data from being present in a merchant or enterprise’s system or network where it could be accessible in the event of a data breach.
Hopefully by using AI and P2PE, merchants and enterprises can stay one step ahead of fraudsters and keep their data safe.
