GS interviews Flywire's Chris Couch about B2B finance

As artificial intelligence reshapes finance and payments, long-standing assumptions about accuracy, workflows and even job roles are shifting at record speed. To explore where B2B finance is headed, The Green Sheet spoke with Chris Couch, head of product, B2B, at Flywire. A veteran technologist and product leader, Couch shared four predictions for how AI will redefine integrations, financial operations, business documents and finance careers by 2030.

Green Sheet: You predict that APIs will become “extinct” as AI agents learn to read. What do you mean by that?

Chris Couch: By the end of the decade, maintaining API integrations will feel as outdated as programming a VCR. AI agents won't require structured data feeds or rigid integration points. Instead, they'll read and interpret documents the way humans do, extracting meaning directly from invoices, contracts, statements and emails.

This shift is already underway. Rather than investing months connecting disparate systems, businesses will deploy agents that instantly understand almost any document format. Every invoice becomes self-service, and every contract becomes self-executing. AI will collapse integration work into simple comprehension.

GS: You also foresee CFOs moving away from obsessing over near-perfect accuracy. How does that change financial operations?

CC: Historically, finance has treated accuracy as the ultimate benchmark. By 2030, that mindset will look antiquated. AI enables companies to operate at high speed even with imperfect accuracy because errors can be identified and corrected instantly.

The economics are clear: an AI that processes 10,000 invoices at 94 percent accuracy and fixes its own mistakes in seconds outperforms human teams that achieve 99-plus percent accuracy but at far lower volume and speed. Smart CFOs are already asking vendors not about accuracy rates but about learning velocity. The real metric becomes the autonomy tax: the tradeoff between correcting AI errors and gaining operational speed.

GS: You've said business documents will "wake up and start thinking." What does that look like in practice?

CC: Within five years, the line between a document and a software program will fade. Invoices, contracts and statements won't just communicate information; they'll enforce actions. An invoice could negotiate payment terms based on cash flow, monitor for duplicates, escalate when ignored, and even update forecasts automatically.

This isn't science fiction. Today's language models already parse agreements and trigger workflows. By 2030, every business document becomes a small autonomous agent. Invoices may email themselves, contracts will flag violations on their own, and financial documents will execute logic without waiting for human review.

GS: You also foresee a major reshaping of finance roles. How will jobs divide in the years ahead?

CC: By 2030, finance jobs will split sharply into two categories: those who train AI and those who handle the decisions AI can't. The middle ground will disappear. Junior accountants won't climb traditional ladders; they'll become AI trainers. Controllers will evolve into risk philosophers. The most valuable professionals will be those capable of encoding exceptions, ethics and edge cases into autonomous systems.

There's no slow transition. As automation accelerates, professionals either teach the machines or focus on strategic judgment. Everyone else gets automated.

Notice to readers: These are archived articles. Contact information, links and other details may be out of date. We regret any inconvenience.