Share with your CIO
Zions Bancorporation is betting that the next competitive gap in banking won’t be opened by generative AI but by agentic AI and the data infrastructure to support it. The bank has partnered with nCino for its core banking software, a deliberate buy-versus-build call, while internally prioritizing three pillars: shifting from content generation to AI that executes tasks autonomously, improving internal operations, and upgrading data quality. No dollar figures are disclosed, but the strategic framing is unusually candid for a regional bank of Zions’ size.
What this means for your business
The institution that decides its primary business is banking, not software development, is already ahead of the ones still arguing the point. Zions’ choice to standardize on nCino rather than build proprietary tooling is the tell here. Banks that have been running bespoke loan-origination platforms built over the last decade are now sitting on technical debt that makes the agentic AI layer (AI that doesn’t just answer questions but completes multi-step workflows autonomously) significantly harder to bolt on. If your organization is still debating build versus buy for foundational financial infrastructure, Zions just answered it by pointing at the collective improvement nCino received from every other bank that adopted it first.
The sequencing Zions describes, generative AI first, agentic AI next, data infrastructure running in parallel, maps a path that most enterprises are handling in the wrong order. The recurring failure mode looks like this: teams deploy a large language model on top of poorly governed data, get inconsistent outputs, lose executive confidence, and stall. Zions is explicitly naming data infrastructure as a precondition, not an afterthought. That’s not a modest operational footnote; it’s the difference between an AI program that scales and one that produces impressive demos and disappointing production results.
The harder question Zions hasn’t fully answered is governance: who is accountable when an autonomous agent makes a credit decision, a customer communication, or a workflow change that a human didn’t explicitly approve? Transparency, which Chantel Chase names as the trust mechanism, is necessary but not sufficient. If your AI roadmap is moving toward agentic systems in 2025 or 2026, the governance architecture for autonomous action, specifically who can halt an agent mid-task and on what authority, is the budget line most likely to be underfunded right now. I’d revise this read if Zions discloses a formal AI accountability structure that sits outside the technology org.
Concept deep-dive: Agentic AI
Agentic AI describes systems that don’t just respond to prompts but pursue multi-step goals autonomously, calling tools, making decisions, and completing tasks without human approval at each step. Think of the difference between asking a calculator a question and hiring an assistant who books the meeting, prepares the brief, and sends the follow-up. In banking, the business implication is significant: the technology shifts from augmenting human judgment to, in some workflows, replacing the loop entirely, which changes both the efficiency math and the compliance exposure.
Based on reporting from Zions turns to nCino as it advances its AI strategy, originally published 2026-10-28 12:07:00.
