In the sprawling universe of enterprise software, artificial intelligence (AI) can often be cast in extreme roles.
Depending on the narrative, AI either propels companies into the future or threatens to displace the very humans who run them.
Yet in a corner of the enterprise often overlooked in the digital revolution, the finance back office, AI is making its mark in more consequential ways.
“You can’t AI everything,” Dave Ruda, vice president of product at Billtrust, told PYMNTS. “There’s no such thing. It’s more, ‘where do we see the most amount of human-in-the-loop manual effort, and how can we make that robotic’ — then eventually move to human-on-the-loop, which requires visibility.”
“You wouldn’t fix a pipe with a hammer,” he said. “You’d use a wrench. AI is just a tool. We deploy it where it clearly reduces friction.”
That friction — often hidden beneath years of manual processes, email threads, spreadsheets and delayed responses — represents a source of inefficiency for enterprise finance teams.
In accounts receivable (AR), for example, where customer communication, payment behavior and risk assessment intersect, latency and lack of insight cost businesses avoidable losses on an annual basis.
Rewiring Accounts Receivable With the Help of AI
For its part, Billtrust has focused on incremental, intelligent upgrades across the AR function. One example is a generative AI-powered email assistant built directly into the collections workflow. The tool drafts responses, surfaces relevant context, and has trimmed average email response time from eight minutes to just two and a half.
The gain is more than just time. It’s reach.
“If a collector can respond faster, they can manage more accounts and spend more time on complex issues,” Ruda said. “It’s not about doing the same work faster. It’s about enabling better work.
“AI is sometimes a shiny object. You touch it once and then toss it away, like a free balloon at the carnival. But this one? People keep using it. It was designed intuitively enough, within the flow of a collector’s normal workday, and it provides value,” he added.
That’s why, in comparison to AI initiatives that chase full automation, Billtrust has drawn a clear boundary between decision support and decision making. The company’s internal development framework prioritizes human-in-the-loop design, ensuring that AI handles the routine while humans stay firmly in control.
That design philosophy has implications not just for product architecture, but for adoption. Collectors, credit analysts and AR professionals — many of whom have built careers on manual expertise — don’t view the platform as a threat. Instead, they see it as an upgrade.
Enterprise adoption is notoriously difficult, especially in departments like finance that have historically lagged in modernization. By presenting AI as augmentation, not automation, finance chiefs can make the leap less intimidating — and more valuable.
“Think of Iron Man,” Ruda said. “Tony Stark is brilliant, but the suit gives him superpowers. That’s what we’re building: systems that scale the intelligence and capacity of our users.”
If there’s one variable that makes or breaks AI adoption, it’s not the model — it’s the mindset. Often the biggest obstacle to innovation isn’t technical. It’s cultural. And while C-level executives may sign the contract, it’s frontline users who determine whether a solution sticks.
“You’re asking someone who may have been doing the same job the same way for 15 years to change,” Ruda said. “That’s hard … [but] people don’t adopt AI because it’s futuristic. They adopt it because it saves them time, helps them succeed, and feels like a natural extension of what they already do.”
Static Limits to Dynamic Intelligence
Where Billtrust moves beyond mere productivity enhancement is in its approach to credit and risk. One of its most powerful — and underappreciated — tools is continuous credit monitoring, an always-on machine learning system that evaluates customer risk in real time.
Traditionally, credit limits are set once during onboarding and rarely revisited. In volatile markets, that’s a liability. A customer who looked healthy six months ago may now be nearing insolvency. Billtrust’s system ingests payment behavior, account activity and external signals to adjust credit exposure.
“We’re treating credit as a living entity,” Ruda said. “Static underwriting is giving way to continuous evaluation. That helps companies be bold — but also smart — about who they do business with.”
The implications are far-reaching. By flagging deteriorating accounts early, businesses can reduce exposure. Conversely, loyal customers who consistently pay on time can be offered expanded credit, driving revenue growth.
It’s a reimagining of finance’s role — from cost center to revenue enabler. And while the momentum is real, so are the challenges. AI is advancing faster than governance. Enterprise finance is notoriously conservative. And models are only as good as the data they’re trained on.
“This isn’t plug-and-play magic. It’s about building systems that are understandable, controllable, and beneficial for the people who use them,” Ruda said.
Cyber Security3 weeks ago
Cyber Security3 weeks ago
Fintech3 weeks ago
Artificial Intelligence3 weeks ago
Fintech3 weeks ago
Latest Tech News3 weeks ago
Fintech3 weeks ago
Fintech2 weeks ago