BNY's payments chief says AI routing is how banks take facilitation back from fintechs

Jennifer Barker, BNY's global head of payments and trade and depository receipts, told American Banker that the bank has invested in AI across its business, including payments, enabling intelligent routing and payments facilitation ^[1]. The competitive claim underneath that is the part worth reading twice: fintechs have offered payments facilitation for years, and AI is what gives banks a new way to compete for the same work ^[2].

Facilitation is not a customer-facing feature. A payments facilitator picks the option for a transaction based on speed, channel or cost, and because faster processing costs more, it sorts a sender's basket of transactions by desired processing time and the fees the sender will pay ^[3]. Barker's framing is that clients do not choose at all; a set of preferences becomes the basis for an automatic decision ^[4]. "Clients aren't concerned about how a payment happens," she said, and do not have to specify real-time, FedNow, RTP or ACH ^[5]. BNY supports this through a single API connection ^[6].

The reason AI enters here is combinatorial, not aspirational. "It's not just selecting the payment rail," Barker said. "There's fraud, smart routing and going through internal controls, as well as pre- and post validation" ^[7]. Add digital assets and the decision space widens again: Barker says a CFO may be choosing between fiat currency, a tokenized deposit or a stablecoin ^[8]. AI, in this account, sifts options against client parameters and routes quickly as those new currency forms appear ^[9].

Note what Barker is not saying. "AI is a capacity creator," she said. "As our business grows you need more capacity to support growth. AI helps us bring more efficiencies" ^[10]. That is an operating-leverage argument aimed at the cost of running more rails, not a revenue story. The one hard number disclosed is modest: in payments operations, AI helps address multicurrency payment exceptions in real time and handles about 10% of payment exceptions ^[11]. Which means roughly 90% still do not go through it ^[12].

Phil Philliou, a payments consultant quoted in the same piece, put the economics plainly: "The beauty of the PayFac model is that banks can monetize sponsorship at scale without owning tech" ^[13]. He describes one use case in which AI and machine learning compress merchant risk review from a manual, multi-system pull into an automated scored package delivered in about 60 seconds, with approved merchants flowing straight into boarding at the processor ^[14]. Underwriting and boarding speed, not routing cleverness, is where sponsorship margins are won or lost.

Sell-side reaction has been favourable. JPMorgan equity analysts said BNY's overall growth has been highlighted by strong growth in its payments and trade units ^[15]. BofA Global Research wrote that near-term tactical execution alongside long-term strategy, including AI adoption and the potential migration of real-world assets on-chain, is what sets BNY apart from many peers ^[16].

What to watch: whether that 10% exception share moves, since capacity claims are testable ^[11]; whether facilitation volumes show up as fee income in payments and trade rather than as a cost story ^[15]; and whether tokenized deposits and stablecoins actually enter the routing set in practice, which is the condition under which a bank's balance sheet becomes an advantage over a fintech facilitator ^[8].