Sunil Rajasekar, CEO of Billtrust, told Karen Webster that while using artificial intelligence (AI) to make better business decisions, generative AI has the potential to change the very nature of payments and commerce.
Billtrust and other firms, he said, have been using AI for a while, harnessed to improve back-office workflows, to set credit limits and to detect fraud.
Generative AI, he said, represents a “step function improvement.”
As he observed, “Things are moving a lot faster than many of us could have imagined.”
GenAI, Rajasekar said, can offer users new ways of engaging with data so that “nuanced conversations” take shape.
The conversation between Rajasekar and Webster was the latest and final installment in the “What’s Next in Payments” series focused on GenAI.
A scaled, measured approach is in order when considering the ways in which AI can improve commerce itself and financial services.
“This isn’t an area where you jump in headfirst,” he said, “You want to understand how the systems work — and use judgment in combination with analysis.”
Browsing 165 Million Books
At its best, generative AI can act like an electronic guide prowling the shelves of the Library of Congress — home to 165 million books, a proxy for the sum total of human knowledge — finding just the right tome, in just the right section, and zeroing on the right page … at just the right time.
“GenAI can get you the answer you want,” he said, “contextualizes those answers, condenses them, adding meaning … and then responding to you.”
That’s the eventuality, anyway, and the ambition.
But much depends, of course, on the data that’s fed into the models, how it’s chosen, with guardrails in place ascertaining that it’s legitimate data.
“If you have bad data, you’re going to get bad results,” Rajasekar cautioned. There is a lot more at stake when relying on GenAI and algorithms to power credit decisioning than there is when offering up browsing recommendations on Amazon.
For payments firms, he said, there’s the benefit of having billions of transactions and years of history at the ready to be extracted from data lakes, with minute details on fund flows, payment modalities and fraud.
“With GenAI,” he said, “you can make up your questions on the fly” — finding out, for instance, who the “best” customers are, what they’ve bought, whether they’ve made timely payments on their cards … and maybe they deserve a card around the holidays thanking them for their loyalty. Billtrust, for its part, has launched an AI “co-pilot,” currently in beta, that helps client firms access and gain insight from their own payments data.
“There are a thousand flowers blooming across the organization when it comes to AI,” Rajasekar said, as the company continues, with a “staggered approach” to develop tools that just about anyone can use.
“Leveraging insights can trigger actions,” he said. He offered up the example of how a trucking company, using AI, can derive insights from macro conditions to uncover demand signals that help determine optimal ways to manage loads and fleets, and whether more trucks need to be on the road, while paying vendors and suppliers and taking in payments from customers.
Hurdling the Headwinds
Yet for all the promise, Rajasekar said, there are still some headwinds in place. Companies are worried about whether their data might be exposed to the public. There still remains the sticky problem of “hallucinations,” where AI models return with incorrect information. In the latter case, he said, the models are becoming more complex and adept at improving “explainability,” and detailing why certain decisions are being made.
Regulation will help ensure there are checks and balances in the AI realm, ensuring that explainability is factored into models as they’re developed and as they evolve. The guardrails and staggered implementations will be especially critical as AI becomes more fully integrated into everything from traffic to defense systems.
Of GenAI’s future, Rajasekar told Webster, “I do believe that the overall trajectory is going to ‘up and to the right.’ There may be stumbles along the way … but I am excited about where we are and what’s coming next.”