Banking’s AI Problem Isn’t The Model. It’s The Plumbing

InnovationBanking’s AI Problem Isn’t The Model. It’s The PlumbingByShuchi Agrawal,Forbes Councils Member.for Forbes Technology CouncilCOUNCIL POSTExpertise from Forbes Councils members, operated under license. Opinions expressed are those of the author. | Membership (fee-based)May 22, 2026, 06:15am EDTShuchi Agrawal: Finance & AI leader "Breaking the Binary" by reimagining legacy banking through human-centric, autonomous systems. gettyEvery bank wants an AI strategy. Far fewer have an AI‑ready bank.​In wholesale and corporate credit, a familiar pattern exists: world‑class people making billion‑dollar decisions on top of brittle systems that were never designed for the speed, complexity and scrutiny they now face. The issue isn’t a lack of models. It’s the architecture they sit on.​Early in my career, leading AI and data initiatives in global banking, I watched deals die in rooms that had nothing to do with pricing or risk. They failed because onboarding systems couldn’t communicate, and APIs didn’t exist. Data sat in siloed platforms, producing multiple sources of truth and no real confidence. Revenue didn’t just slip away; it evaporated.​On the risk side, the story was similar. Teams were talented and well‑intentioned, but the technology couldn’t deliver a complete, timely view of exposure. That’s where you see the downstream effects: delays, remediation work and regulatory findings that trace back not to malice, but to plumbing.​Meanwhile, major industry research now highlights AI and data as core drivers of banking’s next performance curve, from profitability to productivity. The leaders aren’t just piloting models. They’re rebuilding the pipes. From Static Records To Living IntelligenceFor decades, wholesale credit ran on a simple assumption: You could treat data like a file that is captured, stored and retrieved when needed.That assumption is now broken.Markets move faster. Client expectations are higher. Risk events propagate across portfolios an...