Continuous settlement moves treasury's binding constraint from the clock to the ruleset

Three fifths of the pain sits in one line item. Of the 62% of middle market finance executives who told PYMNTS Intelligence and Fynapse, in the August 2026 edition of The 2026 Certainty Project, that they have struggled to manage or scale cash flow forecasting ^[4], 37 points' worth name it their single biggest finance or back-office challenge ^[5], which is about 60% of the strugglers ^[1] and leaves roughly 25 points who find forecasting hard and something else harder ^[2].

That ratio carries the entire working capital argument, because the payoff PYMNTS identifies is less idle cash rather than faster payments, arriving through continuous visibility and programmable liquidity that would let a company hold smaller precautionary balances while still meeting obligations in real time ^[3]. A smaller buffer is only safe to the extent the forecast holds at the hour rather than the quarter.

Citi's Thursday statement names two counterparties and a rail, live U.S. dollar transactions with First Abu Dhabi Bank and OCBC over Swift's blockchain-based ledger ^[1]; SoFi's arrangement with Payward names an always-on schedule for transmitting money and managing liquidity ^[2]; neither account carries a volume, a corridor count, or a fee ^[12]. What is on offer, then, is a capability, and a capability with undisclosed usage is worth whatever buffer it releases, which is a number no one in this material has put on the table.

The governance change PYMNTS describes is that payment policy stops being a static approval matrix and becomes a decision engine ^[6], with treasury software asked to judge whether an instruction falls inside boundaries humans authorized earlier, instead of executing after a signature ^[7]. Existing controls are thresholds bolted to people, two approvals above an amount and extra verification on a new beneficiary ^[8], and the 2am Sunday version of that question is which transactions execute automatically and what happens when the fraud controls fire with nobody at a desk ^[9].

Three ways this plays out differently. Banks could keep an effective cutoff through their own liquidity and credit policy even on a continuous rail, in which case the corporate buffer never moves. Audit and insurance practice could keep a human in the release path for anything material, leaving the engine to low-value traffic. Or the buffer does shrink, and the saving accrues to whoever is pricing the intraday liquidity behind it, which the disclosures here do not identify ^[12].

The version I would defend, or rather the more interesting version, is that the clock was never the expensive constraint: treasury built buffers, approvals and staffing around cutoffs ^[10], and removing the cutoff only converts cash into a policy question that a firm struggling to forecast ^[4] cannot answer cheaply. The test is one ratio in the filings, precautionary cash against monthly outflow. If it sits flat two years into always-on settlement, the buffer was answering something other than the banking day, and this read is wrong.