An S&P Global special report from October 2025 went further, warning that the gap between winners and losers will widen fast. "Banks that secure the benefits of AI—including across costs and revenues—could find themselves with enduring advantages over competitors," the report states, adding that rated entities’ financial and competitive positions are expected to diverge within that same three-to-five-year window.
The early gains are unglamorous but tangible. Shahmir Khaliq, head of services at Citi in New York, points to more efficient treasury management and custody operations as near-term priorities. An AI agent cutting client onboarding time from six months to six weeks—by autonomously checking documents for know-your-customer compliance—is the kind of back-office win that quietly reshapes cost structures.
"We’ll see efficiencies first before we see a lot of visible innovation," says JoAnn Stonier, a former chief data officer at Mastercard who now teaches at Carnegie Mellon University in Pittsburgh.
A sector still shaped by 2008
Banks spent years after the 2008 global financial crisis repairing balance sheets, absorbing new regulations, and coping with record-low interest rates. They also lost ground to private credit and other unregulated providers. The recent cycle of central bank tightening restored some profitability—but now AI is forcing another fundamental rethink of how the industry operates and competes.
85% of use cases are internal—and nearly half of projects collapse
Despite the hype, the AI transformation playing out inside banks is largely invisible to customers. Alexandra Mousavizadeh, co-founder and co-CEO of Evident Insights, a London consultancy that tracks AI in banking, estimates that more than 85% of current use cases are internal. Investment bankers compiling analytical and legal documentation for a proposed merger, or compliance teams automating know-your-customer checks, represent the real frontier—not flashy consumer-facing products.
The failure rate is sobering. Fernandez estimates that nearly half of all AI initiatives within banks fail. "Not every dollar or euro invested results in a solution, and not every solution can scale to where return on investment becomes tangible," she says. Standardizing AI tools across large, data-dense organizations is both laborious and expensive.
"Diffusion takes longer than people think," Mousavizadeh says. "You’re changing habits, not just flipping a switch." She draws a parallel with the internet, whose diffusion across banking took roughly a decade. She expects AI to move about twice as fast—still meaning approximately five years before the technology is broadly embedded.
Data quality is emerging as a foundational competitive divide. "Data readiness, working with data sets that are clean and not duplicated, is a source of competitive advantage now," Fernandez says. Banks that digitized and cleaned their records early are better positioned to feed AI models with reliable inputs.
