Agentic AI Could Make Net 30 Obsolete

It’s the year 2026, and net-30 terms still dominate the corporate landscape.

Companies have digitized invoices, automated accounts payable, connected bank accounts through APIs, and built future-proof, sophisticated cash forecasts. But one of the most consequential financial decisions between businesses, exactly when money changes hands, is still frequently determined by payment terms negotiated months or years earlier.

The common B2B payment term has survived a surprising amount of technological progress. But can traditional payment terms like net 30 and net 60 survive the agentic artificial intelligence era?

If software can continuously understand how much cash a buyer has, how badly a supplier needs liquidity, what either company can earn on that cash elsewhere and what risk each side is willing to tolerate, payment terms stop looking like static contractual provisions.

They start looking like prices.

See also: Working Capital Is Becoming a Priced Portfolio for CFOs

Working Capital Matures From Net 30 to a Continuously Priced Decision

In June, JPMorgan sketched out a future in which a buyer’s treasury agent and a supplier’s treasury agent autonomously negotiate payment timing and discount terms within predetermined policies. The idea is prospective. No current implementation exists, and standardized interfaces, interoperable policies and substantial counterparty trust would all have to come first.

The idea, however, shows what happens when AP starts looking less like a workflow and more like a market.

Consider a buyer holding excess cash on a Tuesday afternoon. Its treasury agent knows the company’s liquidity requirements, short-term investment returns, borrowing costs, cash forecast and supplier risk policies. Meanwhile, a supplier’s agent knows that supplier’s receivables position, financing costs and immediate cash requirements.

Instead of waiting for someone to offer “2/10 net 30,” the systems could theoretically discover a mutually acceptable price for moving payment forward 11 days. That turns payment timing into something resembling a continuously repriced financial instrument. Every approved invoice potentially contains an embedded financing opportunity whose value changes with liquidity, interest rates, counterparty conditions and time.

The “2025-2026 Growth Corporates Working Capital Index,” a Visa report in collaboration with PYMNTS Intelligence, found that 7 in 10 “Adaptive” chief financial officers and treasurers use working capital solutions to pay suppliers faster, stay agile and strengthen supplier relationships in a volatile economy.

Read also: Instant Payments Unlock Working Capital by Allowing Treasury to Pay Later

The CFO’s Autonomy Problem Is Bigger Than AI Alone

The radical part of agentic AP isn’t removing humans from invoice processing. It’s making the economics inside every invoice liquid. Getting there requires solving the problem of knowing where the money is.

The futuristic interface may be an AI negotiating with another AI. Underneath it sits decidedly less futuristic infrastructure built atop ERP records, bank feeds, supplier master data, invoice approvals, cash forecasts and treasury policies that must agree about reality.

Before finance departments can automate negotiation, they therefore must automate certainty. A machine deciding whether paying an invoice 14 days early is economically attractive needs more than an approved invoice. It needs confidence about cash availability, forecasted obligations, the supplier’s identity, contractual restrictions and what alternative return the company could generate from retaining that cash.

The PYMNTS Intelligence Tracker “Who Decides Now: How Developers and Tech Teams Are Reshaping the Future of AP Payments” found in July that 58% of small- to medium-sized businesses (SMBs) rank integration as very or extremely important when evaluating technology solutions.

See also: Everyone Is Selling CFOs Real-Time Finance. Most of It Isn’t.

Corporate treasury policies were designed primarily for people. Agentic treasury requires translating them into rules machines can execute. For working capital, those boundaries could become specific.

An agent might have authority to accelerate payments up to a certain amount when the implied return from an early-payment discount exceeds a hurdle rate. It might be prohibited from extending terms for strategically critical suppliers. Another policy could require human approval whenever negotiations materially alter expected monthly liquidity.

The distinction is important because payment timing touches liquidity, supplier relationships, procurement strategy, credit risk and potentially accounting and contractual obligations simultaneously. The agent therefore needs something closer to a financial constitution than a chatbot prompt.

None of this means CFOs are about to turn AP over to autonomous agents. The near-term evolution is likely to move from AI identifying working capital opportunities to recommending them, to executing narrowly defined decisions and only eventually to negotiating with external agents.

But the destination reveals something about what today’s automation projects may actually be building toward. The goal is no longer merely faster AP. It is an environment in which every payable can potentially become a real-time decision about the highest-value use of corporate cash.

For all PYMNTS B2B and AI coverage, subscribe to the daily B2B and AI newsletters.