# AI in Banking: The Dual Workforce Is Already Here
_Published: 2026-06-30T10:00:00.000-04:00_

_Banking leaders have stopped debating whether their workforce will combine people and AI. New nCino research shows they are already building it._

In the 1970s, the ATM arrived, and the predictions came fast. Skeptics thoughts the bank teller was finished. Why keep a person to count cash and take deposits when a machine could do it faster, and around the clock? The teller, surely, would be the first casualty of banking automation.

But that didn’t happen. Machines handled the routine transactions, which made branches cheaper to run, so banks opened more of them, and teller employment continued to climb through the rollout. The job didn’t disappear; it changed. It moved towards the customer.

You may have heard this analogy lately as banking turns its attention to artificial intelligence (AI). It's the story the banking industry defaults to whenever a new technology threatens a role. AI, like the ATM, can hand people back the hours that routine work consumes and free them for the parts of the job that carry high value — the relationships, advice and judgment.

Part of that reassurance holds true, but it misses a key factor: **speed. **The ATM shift took decades, and that pace was much of the reason it went well. Banks, tellers, and customers had time to adjust as it happened. AI is arriving in a fraction of that time, which is why this shift can't be left to sort itself out.

This time around, leaders aren't waiting to see how it ends. In the [nCino AI in Banking Benchmark,](https://www.ncino.com/ai-in-banking-benchmark) a survey of 150 U.S. senior banking executives, **89% said their organization will be a combination of AI agents and human employees within five years — a dual workforce.** This shift has stopped being a forecast. Leaders are already deciding how to staff it, fund it, and govern it.

In their breakout session at [nSight 2026](https://explore.ncino.com/nsight-2026?_gl=1*1rkl6zq*_gcl_aw*R0NMLjE3ODE3MDcyOTUuQ2owS0NRandpOG5SQmhEaEFSSXNBSFpmX3Bia2x3QVp5S1k2bXZTZktoUmlnQ1JUNzQtMUt6Q0MwVXFMdml1X2tHaHBONGpaanRpUE83RWFBb0ZSRUFMd193Y0I.*_gcl_au*NTg0MDU4MTI0LjE3ODIxNTQ2MzY.*_ga*NTQ2NjI1MTAzLjE3NzQyNzY1MzI.*_ga_F72SC3LCNT*czE3ODIyMjE3NDckbzE5NSRnMSR0MTc4MjIyNDEyMCRqNTAkbDAkaDA.), John Fimbel, Executive in Residence at Duke University, nCino CTO Will Jung, and Jim Marous, Co-Publisher of The Financial Brand, worked through what a dual workforce really looks like and what it requires. The benchmark data and that conversation at nSight point to the same conclusion: the technology question is mostly settled. The harder questions are about people, accountability and data.

### **What a Dual Workforce Actually Means**

A **dual workforce** pairs people and AI agents around the same work, with each handling what it does best. Just as ATMs did in the 80s, machines take the repetitive, high-volume tasks and people keep the work that depends on judgment and relationships.

Fimbel described it as the overlap between experienced professionals, entry-level talent, and AI working alongside both. As Jung put it, AI is a tireless pattern matcher that runs around the clock, learning from massive volumes of data that no person could. People hold the judgment, the relationships, and the accountability.

Picture what that looks like in practice. A commercial banker walks into a client meeting with AI having already pulled the relevant data from inside and outside the bank, ready to solve a problem before the client raises it. The machine does the gathering; the banker builds and nurtures the relationship. During the breakout, Marous described that opportunity from the banker's side of the desk.

> "If I was a client officer, it would be so exciting to take all these AI tools, bring the data to the client I'm trying to work with, help solve their problems, and be the hero in their eyes."
> 
> — Jim Marous, Co-Publisher, The Financial Brand

There are some parts of the work that can never move to the machine, no matter how capable it gets. As Jung told the room, a computer can never be accountable, so a banker always will be. Responsibility for a lending decision, a customer relationship, and a regulatory obligation always stays with a person.

That’s how we think about it at nCino. **AI is an operating model shift rather than a tool that replaces bankers’ judgment. **Our [Digital Partners,](https://www.ncino.com/blog/digital-partners-redefining-banking-teams) a set of AI agents aligned to specific banking roles, take on defined work inside a banker’s day, while our conversational interface [Banking Advisor](https://www.ncino.com/banking-advisor) brings answers and context into the workflow. The goal is to give teams back the hours that routine tasks consume, so they can spend more time on their relationships with their customers and members — where banks can actually differentiate.

### **The Shift Is Already Underway**

With the ATM, the workforce settled into a new balance over decades, mostly by accident. The banking industry understood what had happened only in hindsight. But the AI transformation is already underway, and banks are making decisions in real time. According to the benchmark, 84% of executives are AI users today, and 84% said agentic AI has already changed how banking roles operate. Ninety-one percent said AI is freeing their employees to focus on higher-value, customer-facing work.

While we don’t have decades to adjust, we do have historical context that the banks in the 70s and 80s didn’t. We have the choice to move intentionally, and it starts with our people.

Leaders are already reshaping their org charts around the AI era. To prepare their workforce, 55% of surveyed senior banking executives say they’re reskilling current employees, 48% are adding digital and technology roles, and 42% are creating new AI-focused positions.

In the 1970s, no one sat down and decided what the bank teller’s role would become once ATMs rolled out. The role just formed, slowly, as branches multiplied. This time, the dual workforce is something financial institutions can design deliberately. That's the opportunity.

### **The Shift Is Moving Faster This Time**

Speed is what makes this shift different from the ones banking has weathered before. Fimbel walked through innovation cycles we've already gone through — the loom, the automobile, the spreadsheet, the internet — and made the point that each one arrived faster than the last. AI is the steepest jump yet. As Jung noted, agents only became capable of working on their own within the past year. The decades-long rollout the bank teller had doesn't exist this time.

This changes the reward for moving early. When a shift takes decades, you can hang back, watch the early movers, learn from their stumbles, and still catch up. At the speed of AI, the institutions that move first set the pace. Marous argued that fast following is no longer a safe place to be. The benchmark shows leaders already adjusting to that reality, with 81% saying their organizations are prioritizing AI adoption ahead of proven return on investment. That willingness to act before the returns are fully proven suggests leaders aren't treating this as a wait-and-see investment.

### **The Data Underneath It All**

A dual workforce works when the AI half can be trusted, and that comes down to data. The benchmark shows both the confidence and the work ahead: 87% of executives said they are confident in the quality of their data, while 93% reported at least one data governance challenge. The most common was siloed data, cited by 52%, followed by concerns about data integrity and consistency. An integrated system is what turns that data into something AI can act on, likely why 94% said they want AI delivered as one end-to-end capability rather than a patchwork of point solutions.

Speeds rewards getting this right. An agent acts on whatever data it has at any given moment, so the foundation needs to be sound first. Institutions investing in that foundation now will move the fastest. As Jung told the nSight audience, **AI is only as good as the context it can draw on. **Governance often is seen as an anchor, he said, when it can be a differentiator. When risk management gets easier for internal teams, that ease shows up in the customer experience.** **

This is where banking-specific context becomes a real advantage. nCino's platform-native AI draws on more than 14 years of banking workflows and the anonymized data context of more than 1,500 customers, so it understands the work it supports. Generic models understand language. Banking-specific AI already understands lending, risk, and the way a real institution operates. Because that context is built in, an institution can keep pace instead of building it from scratch.

### **Where To Start**

With the dual workforce arriving faster than shifts before it, the hardest part can be knowing where to begin. Fimbel, Jung, and Marous offered some practical advice.

- **Begin where the friction is highest. **Think about the workflows that consume the most time but require the least amount of judgment and let AI take the first pass. In the benchmark, executives are already leaning on AI for work just like this, from summarizing financial documents to supporting credit analysis.
- **Show early wins. **Building your team's confidence in what AI can do is as much about change management as technology, so demonstrate value early instead of waiting for one grand reveal.
- **Pair people with agents intentionally.** Keep a clear view of which decisions stay human, so the pairing is built by design.

Fimbel ended the session on a more personal note, with a reframe for anyone wondering where they fit. Instead of asking _what happens_ if AI takes your job, ask how it _could_ take your job. Once you know which parts of your work AI can handle, you can spend more of your time on the parts it can't. As he put it, many of our jobs will pivot over time in the dual workforce environment. The people who make that pivot are the ones who will come out ahead.

### **Building Your Dual Workforce with Intention**

In the ATM era, the teller's role moved toward the customer, and banks had decades to adjust to it. The benchmark says the same kind of shift is here again, with nearly nine in 10 leaders expecting a dual workforce within five years. This time, it's happening on a fraction of the timeline. That difference is where the opportunity lies. The work ahead is to pair people and AI with intent, keep accountability and relationships firmly in human hands, and get the data in shape to support both. Institutions that do this will be more equipped to move at the rapid speed of change.

At nCino, that's the future we are building with our customers, where agentic banking and human expertise reinforce each other. [The leaders and institutions that design for that pairing now](https://www.adammendler.com/blog/ai-and-the-workforce/) will set the standard the rest of the industry follows.

Find the complete data behind these findings, plus the conversations and solutions shaping the dual workforce at nCino, on the [nCino AI in Banking Benchmark](https://www.ncino.com/ai-in-banking-benchmark) page.

---

[View sitemap](https://www.ncino.com/api/markdown/sitemap)