Adoption was never going to be the hard part. Most banks have already brought AI into the building. The harder question, and the one still separating leaders from laggards, is how to turn that adoption into results that show up on the balance sheet.
The nCino AI in Banking Benchmark lays the gap bare: 84% of banking executives say AI has already significantly changed how most roles operate, yet just 21% can draw a line from that investment to revenue. Closing that distance is arguably the clearest opportunity in banking today, and the institutions making progress are following the same sequence: trust first, then time, then outcomes.
Trust is the foundation, not an afterthought. Banks making real progress are mapping their data, setting clear rules on who can use it and defining how automated decisions get reviewed before an agent goes live. That groundwork lets AI become a source of advantage rather than risk, particularly as customers, regulators, employees and boards all watch for the same thing: an institution that can explain what its technology does and why.
The World Economic Forum reinforced that point in an April 2026 board-level playbook on governing agentic AI, calling for “legible friction”, deliberate pause points where a person signs off on high-stakes actions, and making clear that liability cannot be outsourced to the machine.
Once that trust is built, time follows. Agentic AI is increasingly handling the assembly work, pulling documents, keying figures and chasing missing statements, freeing analysts, relationship managers, loan processors and portfolio teams to focus on judgement rather than admin. McKinsey’s 2025 Global Banking Annual Review describes a near-term model where a single employee oversees 20 to 30 AI agents running end-to-end workflows, effectively creating a dual workforce of people and agents working in tandem.
Time returned is only valuable if banks commit to using it, and a handful of institutions are proving what that looks like. ConnectOne Bank has cut a 20-minute document task down to seconds through embedded AI across commercial lending. ConnectOne Bank chairman and CEO Frank Sorrentino III said, “…I believe with the things we’re working on today together with nCino, we are going to be able to make every single one of our frontline people 50 percent more efficient. Fifty percent means our bankers will work a thousand hours less on things that don’t matter and a thousand hours more on the things that do.”
UK lender ThinCats has saved around 25 hours a month by automating financial covenant testing, while Bendigo Bank consolidated more than 30 forms and systems into a single platform in just 13 months, one of the fastest transformations of its scale across APAC.
The Cambridge Centre for Alternative Finance found in its 2026 study that 55% of financial services firms struggle to measure the value of their AI deployments, rising to 76% among large institutions, which is exactly why these results stand out. The banks getting there aren’t necessarily running better technology. They’re choosing to commit rather than pilot, while the advantage is still there to capture.
For nCino, the message is clear: AI’s next phase in banking will be measured less by adoption and more by whether institutions can build the trust and workflows needed to turn it into measurable returns.
Copyright © 2026 FinTech Global
