Green Sheet interviews Paystand's Scott Bennion

As artificial intelligence becomes increasingly capable of automating finance operations, many organizations are discovering that traditional payment infrastructure is limiting how far that automation can go. At the same time, stablecoins are gaining attention as a potential foundation for faster, more data-rich enterprise payments.

In this Q&A, Scott Bennion, chief financial officer at Paystand, discusses why finance leaders are taking a fresh look at payment rails, the challenges of integrating stablecoins into existing workflows and how programmable payments could reshape accounts payable, accounts receivable and cross-border transactions.

Green Sheet: 1. As finance departments adopt agentic AI, where are traditional payment rails creating the biggest operational bottlenecks today?

Scott Bennion: The bottlenecks show up in three specific places, all connected by the same root cause: legacy rails move money but not the data that explains why it moved.

In accounts receivable, payments arrive via ACH or wire with minimal remittance information — no invoice number, sometimes just a lump sum covering several invoices at once. Your AR team manually researches each payment, matches it to open invoices and posts it to the ERP. That reconciliation work happens after every transaction and gets slower as your customer count grows. It's why companies on legacy rails routinely spend the first half of every month cleaning up the last month's payments, with close cycles stretching to 14 days or more.

On the AP side, an invoice arrives, gets manually coded to a general ledger account, routes through an email-based approval chain, and then gets re-entered into the payment system as a separate step. Each handoff is a delay, and each delay is a cost.

International payments compound both problems. Wires pass through correspondent banking chains with multi-day settlement and little visibility in transit. Treasury teams end up carrying excess cash as a buffer against that uncertainty—capital sitting idle when it should be working.

What makes this especially costly right now is that agentic AI is ready to handle most of this work automatically: prioritizing collections, coding invoices on arrival, running compliance screening before payments release. But the moment AI needs to actually move money, it hits infrastructure that runs on batch cycles and banking hours.

The AI can flag a past-due invoice at 11 p.m .Sunday; the payment still waits until Tuesday. According to McKinsey, modernizing these workflows could reduce process costs by 30 percent to 50 percent, but only if the payment infrastructure underneath actually supports automation. The bottleneck was never the AI. It's the infrastructure the AI is forced to work around.

GS: 2. Why are stablecoins increasingly being viewed as infrastructure for enterprise payments rather than simply as cryptocurrency investments?

SB: Because the actual usage data no longer supports the "crypto" narrative. According to McKinsey and Artemis Analytics, B2B payments now represent roughly 60 percent of real stablecoin payment volume—and that B2B segment grew 733 percent year over year in 2025. That's not traders speculating but rather businesses paying suppliers, contractors and counterparties through a faster, cheaper channel than the one their bank offers.

The practical case is straightforward. A wire transfer to Mexico costs 2.5 percent to 6 percent in fees, takes three to five days, and passes through multiple correspondent banks that each introduce delay and opacity. A stablecoin payment settles in minutes at a flat rate with a full audit trail. For a company running global operations, that difference shows up directly on the cost line and in how much idle cash treasury has to hold against settlement uncertainty.

What's changed structurally is that the institutional plumbing is now in place. The GENIUS Act in the U.S. established the first federal regulatory framework for payment stablecoins. Stripe acquired Bridge for $1.1 billion to build stablecoin payment infrastructure. Mastercard acquired BVNK for $1.8 billion for the same reason.

With actual stablecoin payments in 2025 totaling close to $400 billion, the question stops being "is this legitimate?" and starts being "how do we operationalize it?"

The gap that remains—and what we're addressing with USDb—is that the commonly used USDT and USDC were built for crypto markets, not for enterprise workflows. They move value but don't carry the invoice data, approval context or ERP connectivity that a finance team actually needs. That's the infrastructure layer still being built.

GS: 3. The "last mile" problem appears to be slowing stablecoin adoption among CFOs. What specifically makes integration into existing finance workflows so challenging?

SB: The "last mile" metaphor implies stablecoins are almost there—just a short distance from being useful. We don't think that captures the actual problem. The more real framing is that payment infrastructure was built to move money, and enterprise finance was built to track it. Those two systems have never been designed to talk to each other, and we need a shift to new rails to make this happen.

Here's what the challenge looks like in practice. A company issues an invoice. It contains the vendor name, the purchase order number, the GL code, the approval chain, the contract terms, the compliance flags—all the context a finance team needs. The moment payment initiates over legacy rails, almost all of that context disappears.

An ACH transfer arrives as a dollar amount from a bank account. Someone on your team now has to reconstruct what it was for, match it to the right open invoice and manually post it in NetSuite or SAP. That work happens after every single transaction and grows linearly with volume.

USDC and USDT don't solve this problem. They move value faster and cheaper, but they weren't designed to carry invoice data, trigger ERP reconciliation or enforce approval workflows on settlement.

What's required isn't just a faster payment but a settlement layer that arrives with the business records attached. That's the missing fabric that USDb is designed to provide: money that moves with enough context that the ERP can process it automatically, without a human reconstructing meaning after the fact. Speed matters, but connected settlement is the actual unlock.

GS: 4. How could programmable payment rails change processes such as accounts receivable, accounts payable and cross-border payroll in practical terms?

SB: "Programmable" simply means you can attach rules to money. A payment isn't just an instruction to move funds. It carries conditions. Release this payment when the purchase order is approved. Apply this cash to invoice 12345 at settlement. Lock in the FX rate at authorization, not three days later when the wire clears.

In accounts receivable, that eliminates the manual matching work that currently happens after every payment. Instead of an amount arriving from a bank account with no context, cash posts directly to the right invoice in the ERP at settlement. In one case, that change recently cut days sales outstanding by 80 percent and per-transaction fees by 98 percent. That's not an efficiency gain. It's a different cost structure entirely.

In accounts payable, the bigger shift is where control lives. Today, most AP teams approve an invoice, then execute the payment, then reconcile it—three steps, often handled by different people across several days.

With AI agents on programmable rails, those collapse into one motion. The agent reads the invoice, assigns the GL code, routes approval through Slack or Teams, and when approval lands, the payment executes and posts to the ERP simultaneously. Policy enforcement moves from after the fact to before the transaction.

Cross-border payroll offers the best demonstration of what's possible. Traditional international wires require pre-funded accounts in each country, absorb significant fees and take days to settle. Stablecoin rails replace that with a single USD balance, same-day settlement in 190+ countries at a flat rate, with sanctions screening and dual-approval controls built into the process.

For any company with a global workforce or supply chain, that's not incremental. It changes the economics of operating internationally.

GS: 5. Many finance leaders remain cautious about stablecoins because of compliance, governance and auditability concerns. How are those concerns evolving?

SB: The concerns are legitimate, and I'd worry about any CFO who dismissed them. You can't run a serious enterprise on infrastructure your auditors don't understand and your lawyers can't sign off on. Before last year, the stable coin regulatory picture was genuinely murky: no clear anti money laundering framework, no sanctions guidance and a blockchain audit trail that most ERP systems couldn't consume. Finance leaders were right to wait.

What's changed is the regulatory foundation is now visible. The GENIUS Act, signed in 2025, establishes that permitted payment stablecoin issuers are treated as regulated financial institutions for money laundering and sanctions purposes—the same compliance obligations that govern banks.

Europe and Asia have created similar regulatory structures. CFOs can now have a concrete conversation with their legal and compliance teams and get a real answer instead of hypotheticals. Auditors and regulators are also simply more familiar with the structure than they were two years ago, which has practical value when you're trying to get a new payment method past your audit committee.

On auditability, blockchain-based payments are more transparent than traditional wires in several respects: every transaction is time-stamped, immutable and traceable. The missing link was making that trail legible inside the systems finance teams already use. That's an integration problem, not a technology limitation. We build settlement-to-ERP reconciliation into our platform so every stablecoin transaction produces a record the customer's auditors can actually work with.

Governance is ultimately a design question: are approval workflows, dual controls and policy enforcement built into the payment process, or bolted on afterward? For us it's the former—because a CFO who can't show their board a clean control environment won't adopt anything, regardless of the cost savings on the table.

GS 6. You've suggested that agentic AI could accelerate stablecoin adoption in the $100 trillion B2B economy. What would widespread adoption actually look like over the next few years?

SB: The honest answer is that it won't look dramatic from the outside. It won't be a moment where businesses announce they've "switched to stablecoins." What it will look like is a quiet but significant shift in where efficiency gains come from.

The companies that get there first will report that their finance teams are doing more work with the same headcount. Their close cycles will shrink and their cost to move money will drop. From the outside, it will just look like they're running better. From the inside, it's because their financial operations are increasingly automated and running on rails that actually support that automation.

The near-term adoption path, over the next two to three years, runs through workflows where the pain is most obvious and the ROI is easiest to quantify. Cross-border payments is at the top of that list—any mid-size company paying suppliers across multiple countries has an immediate cost and speed case that's straightforward to measure.

Global contractor payroll follows closely. Accounts receivable automation comes next, because it connects directly to cash flow, which every board tracks.

The longer-term picture is that agentic AI and programmable rails become co-dependent. AI agents need to move money, and the only infrastructure that supports machine-speed, machine-initiated transactions is digital rails. As AI handles more routine finance work—collections, approvals, reconciliation, vendor payments—the volume of machine-initiated transactions grows, and legacy rails become an increasingly poor fit for how the work actually runs.

Chainalysis projects stablecoin transaction volume could reach at least $719 trillion by 2035 on current trajectory. We think B2B commerce is where a significant portion of that lands. The infrastructure is ready, the regulation is clearing and the AI systems that need these rails are already being deployed. The window between early adopter and industry standard in B2B payments tends to be shorter than people expect once the fundamentals align—and they're aligning now.

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