Scaling AR without scaling the team: How GPRS Put AI Agents to Work

Every finance leader has heard the agentic AI pitch. Far fewer have seen it running inside a real AR team, on real volume, with a Controller accountable for the numbers.

GPRS, the nation's largest utility-locating and concrete-scanning company, is growing 15%+ a year. That growth lands on AR as 40,000+ invoices a year, five buyer portals, and a 15-person team. Static aging reports and rule-based automation were already in place. They weren't moving the needle: collections were worked invoice by invoice, cash was applied by hand, and nobody had a live view of the cash position.

So GPRS changed the playbook, not just the tools. Collections moved from region-based to priority-based. The biggest balances and highest risks go first. And AI agents took over the growing work that kept requiring more FTEs, following up on every email thread, matching payments, and submitting invoices across Coupa, Ariba, and Tungsten.

The result: invoice volume up 35%, with the same AR team.

In this session, GPRS's Controller Andy Riggs walks through what that looks like day-to-day: what the agents run on their own, where his team still makes the call, and what changed for the people doing the work. No hypotheticals, no roadmap slides. Just how one high-volume AR team runs today.
‍

What we'll cover

• Why rule-based automation stopped scaling at 40,000 invoices a year
• The playbook shift: region-based → account-based; manual → agentic, etc
• The agents in action: AR inbox, cash application, and buyer-portal submissions
• Where GPRS lets agents act, and where the team still makes the call
• What changed in the numbers, and how fast