Sudhir Jha, senior vice president and head of Mastercard’s Brighterion unit, told Karen Webster in the most recent On the Agenda discussion that artificial intelligence (AI) can strengthen credit and risk management and broaden its value well beyond simply improving day-to-day operations.
But to get there, enterprises need a bit of guidance.
“What used to be cutting-edge technology five years ago is no longer cutting edge,” he said, and enterprises that try to keep up with the rapid changes in data science and analysis on their own can be quickly overwhelmed. The enterprise that starts with regression and pattern analysis solutions might scale rapidly and find benefit from neural networks.
For banks, acquirers and healthcare payments executives, he said, using vendors’ AI-based solutions help to avoid undue losses from fraud, the abuse and misallocation of funds and poor underwriting decisions.
Jha’s comments came against a backdrop in which 88% of financial institutions (FIs) said the pandemic has made lending and credit more challenging. Jha said that the pandemic has underlined that firms need fraud fighting systems in place that are adaptable and that change as the fraudsters change.
“Every month, even every day can be different,” he said, adding that “as an enterprise, there is only so much that you can react to.”
Although most executives see the value of deploying AI systems in their risk-management efforts, many executives are unsure of just how to proceed in making the aspiration a reality. AI tech talent is hard to find and takes years to train, and building models and solutions in-house can take years.
Given those challenges, it makes sense to work with outsourced providers (Brighterion among them), where tapping into what Jha termed a fully baked solution can be no different than the way buying an HR solution a decade ago might have been.
We’re moving toward turnkey solutions where AI modeling is baked in, with the different, disparate data elements that firms need to boost their own fraud defenses.
“Your own data gets into the system, and as the model trains on that, you’ll get more and more value — and you can get started pretty quickly,” he said. “That’s the next evolution for AI solutions.”
Along the way, rapid, accurate decisioning allows FIs (and other enterprises) to build new credit products and services — evident in the explosion in buy now, pay later (BNPL) offerings in just the past year across traditional FIs and digital-only FinTechs, he said.
Improving Credit Risk Management
Jha noted that AI has particular value in managing credit risk. It can leverage real-time data to help lenders make better decisions at origination and spot hiccups and potential fraud before losses hit results.
There’s a bit of lopsided embrace of AI as 79% of banks with more than $100 billion in assets use AI, but only a fraction of smaller banks do. And although progress has been made, the greenfield opportunity is significant. In 2018, 5% of FIs reported using AI systems in areas like credit risk management and fraud detection. By 2021, that figure had increased threefold to 16%.
Jha posited that smaller firms might not have enough data (particularly as the initial credit application is made) to build AI models. Even on the delinquency side of the equation, there is not enough data on hand to be predictive. He said that platform models have broader options that enterprise clients can access to build the best models possible.
The urgency to tap into the platform model is there, as 93% of all acquirers said they are seeing more fraudulent transactions than one year ago. Ninety-eight percent of all acquirers that use AI use it to detect fraud, and 79% of them consider AI to be the most important tool in their fight against fraud.
“Everybody’s struggling with onboarding hundreds and thousands of small businesses that they want to do business with,” he said.
The Approval Dilemma
Acquirers face a dilemma when fighting fraud, said Jha. On one hand, they want to make sure that the fraud is stopped early, before those suspect transactions are submitted to the issuers. But they also want to increase approval rates — in Jha’s terms, “because the issuer knows you’re a good acquirer.”
But acquirers must also worry about the risk of onboarding fraudulent merchants. Machine learning (ML) and AI technologies can accelerate the process. He told Webster that Brighterion is bringing its data assets together to create a data onboarding solution that can assess credit risk at that initial point of contact.
“You’re building more safety into the system,” he said, as each transaction and customer is viewed holistically.
Looking ahead, Jha said that we’ll see a continued shift away from enterprises continuing to invest in AI by building solutions from scratch.
“You’re not actually going to get any differentiation by doing that,” he said. “If you can actually have a vendor that can provide you a turnkey solution and will invest in innovation going forward, you are much better off just adopting that, and then making some custom improvement on that.”
As he told Webster, in credit risk and anti-fraud efforts, “adaptability has become the name of the game.”