How AI’s Language-Based Capabilities Will Transform B2B Payment Processes

Over a long enough time, change stands alone as the only constant.

After all, it was over a century and a half ago that Charles Darwin first wrote that “the species most responsive to change” is the one that survives — and while he was talking about evolutionary biology, the truth captured is universal. That principle can be equally applied to B2B payments as to the finches of the Galapagos.

That’s particularly as it relates to the potential of artificial intelligence (AI) in optimizing accounts receivable (AR) processes.

“You can look at AI in two lenses, internal efficiencies and external product and then break that down into where the low-hanging fruit is,” Ahsan Shah, senior vice president of analytics and AI at Billtrust, told PYMNTS.

Traditionally considered back-office functions, both AR and accounts payables (AP) are now evolving into strategic components of business operations — and areas ripe for AI integration.

“When I look at payables versus receivables, they are two sides of the same coin,” Shah said, highlighting the importance of looking at the entire lifecycle of receivables as a system, involving buyers, suppliers and processes.

Utilizing generative AI in various areas such as deciphering unstructured text, resolving disputes, automating processes and personalizing recommendations all represent some of the “low-hanging fruit” where AI can have an immediate impact, he added.

Leveraging AI-Driven Workflow Enhancements Across B2B Payments

The ability of language models to understand and interpret data, whether it’s invoices, receipts or communications, marks a significant shift toward making AR a strategic asset for enterprises.

Shah said Billtrust is applying generative AI to its own datasets to gain insights into payment trends, buyer behavior, risk analysis and anomaly detection.

“Historically, it has been a struggle trying to get this data aggregated so that we can make decisions,” he said, emphasizing that generative AI helps solve existing problems in a “more accelerated” way, as well as helping to further automate processes and improve analytics for better decision-making.

In the broader B2B payments ecosystem, nearly every business has various needs that can be addressed through generative AI.

Automation and integration of payment systems, handling unstructured data from communication channels, language translation for cross-border payments and personalized recommendations for dispute resolution are some of the areas where AI can provide accelerated solutions, Shah said.

“We’re going away from a rule-based universe to a language-based universe, which opens the doors to many, many possibilities within payments,” he added, explaining that generative AI can give engineering teams and internal code a shot in the arm, making enterprise resource planning and other software integrations more efficient.

Deploying AI

“The first step when undertaking an AI integration is to not shy away from the technology,” Shah said.

He emphasized the need for organizations to prepare for AI integrations by ensuring data readiness. This involves understanding the differences between structured and unstructured data and adopting suitable database solutions like vector databases. The quality of data plays a critical role, and organizations should strive to clean, tag and structure their data effectively.

Despite the hype surrounding AI, Shah advised going beyond surface-level discussions and exploring how AI can be applied in specific contexts.

“When people ask me where generative AI can be used, my answer is where can it not be used?” Shah said. “But there’s going to be a stepwise function going from the basic generative AI to something that’s really going to differentiate your business … Step one is to understand which problems you’re going to solve.”

By starting with a plan and a roadmap, organizations can gradually evolve their AI adoption over time, keeping pace with the rapid advancements in AI research and development.

As Shah said, “AI is only as good as the data and the enterprise data infrastructure that’s available … But I’ll make the bold statement that AI is essential for every organization.”

Looking ahead, Shah predicted a shift toward agentic, or multi-agent generative AI. This involves a layer above individual processes, where AI acts as an orchestrator across various domains and can carry out tasks rather than just surface information.

This approach will optimize various steps in the AR process and enable personalized recommendations, campaign management and strategic decision-making. The focus will shift from point solutions to a holistic approach that maximizes the potential of generative AI.

“This is a transformational opportunity for B2B,” Shah said.