Editor’s Note — This article is sponsored by Chime. As with all sponsored content in Fintech Takes, this article was written, edited, and published by me, Alex Johnson. I hope you enjoy it!
Here’s a question I’ve become fixated on lately: What, exactly, is a primary bank account?
Mechanically, the answer is straightforward. It’s the demand deposit account a consumer uses most often — the one where the paycheck lands and the bills get paid. The operational center of their financial life.
But there’s a second definition, at the relationship level, that’s simpler and more profound. It’s the answer to a question you could pose to any consumer:
Who do you bank with?
I like this relationship-oriented definition because it evolves with the market. Trace it and you can see banking primacy move through three distinct eras — with a different answer to what actually earns a bank the relationship: first, physical proximity; then, digital convenience; and now (I would argue) trust. That third shift is the one I think matters most. As agentic AI works its way deeper into how people manage money, trust stops being a nice-to-have and becomes the whole game — but more on that in a second.
30 years ago, primacy in banking was a function of distribution. You banked with whichever bank was most convenient, and convenience was defined by physical proximity: how many branches did they have in your neighborhood? Distribution was destiny. Call this the first era of primacy: Built around the proximity to a banker at a physical branch who – at least in theory – knew your name, financial history, and preferences. Depending on where you lived, you might get to choose from a handful of banks. Or there might literally have been just one.
Today, the internet has rendered that distribution advantage irrelevant. Every bank account is equally convenient because every bank account is one tap away on your smartphone. This is the second era: Digital self-service; the currency for banks is no longer physical reach or branch size, but convenience.
This shift raises an uncomfortable question: Does the concept of a primary bank account still make sense in 2026?
Mechanically, the answer is sorta. Consumers will likely still have one primary operating account that most of their money flows through, though, given the ease of account opening and money movement today, it’s not strictly necessary like it used to be.
But at the relationship level? Do consumers still have a clear and unimpeachable answer to the question, Who do you bank with? Does that question even make sense anymore?
There’s evidence that it may not.
J.D. Power’s 2026 U.S. Retail Banking Satisfaction Study found that the average retail bank checking customer now maintains three deposit accounts at different institutions, and 20% of customers have moved money away from their primary bank within the past three months, up from 17% last year. The customers most likely to be moving money? Those under 40 (23%), the affluent and mass affluent (25%), and the financially healthy (24%). In other words, the customers banks most want to keep.
Consumers seem increasingly unsure who they want to bank with. They’re keeping multiple deposit accounts open while they figure out which account, or combination of accounts, they want to consolidate on.
And here’s why this moment feels genuinely precarious for banks: It’s possible consumers won’t choose to centralize their banking activity with any one institution at all.
Instead, they may centralize the relationship (who do you bank with?) while keeping the underlying accounts as decentralized as they are today. Or more so.
In theory, the technology that could facilitate this outcome is agentic AI — an intelligent, automated system that interacts with the customer on the front end and orchestrates a complex web of accounts and workflows on the back end. In this world, the answer to “who do you bank with?” becomes something like, “I don’t really know where all my money is stored, but I do all my banking through ChatGPT.”
We’d go from a Primary Bank Account to a Primary Banking Agent.
And we’re already moving in this direction. OpenAI has explicitly positioned ChatGPT as a financial agent and says that more 200 million people a month are already asking it finance questions, although I suspect the majority of these questions are basic financial literacy, research, and advice.
This is the democratized version of a concept that’s long existed in wealth management: The deposit broker. Rich people have always been able to hire someone to shop their cash across institutions, chasing yield with zero emotional attachment. A non-bank AI agent would do the same thing for everyone — managing your money ruthlessly, with no loyalty to any bank and no net interest margin of its own to protect. Better rate somewhere else? Consider your deposits moved. You’d know who holds your money the way a homeowner knows who services their mortgage, which is to say, you wouldn’t.
This notion is appealing in a lot of ways. But there’s a problem with it: Accountability.
Abraham Maslow’s Hierarchy of Needs is a useful analogy here.
We valorize the top of the pyramid — self-actualization — and forget that it’s only possible on top of sturdy foundations of safety and security. You don’t optimize your career while your house is on fire. In financial services, the foundation translates to: Does my money work? Is it safe? Is someone accountable if it isn’t?
At the end of the day, consumers want their money to be safe. They want someone they can talk to — and take action — when there’s a problem. They want someone who is accountable to them. Someone whose interests are structurally aligned with theirs.
In short, they want someone they can trust and act with their permission.
AI labs like OpenAI want to own the interface and customer interaction layer, but I think it’s unlikely that they’ll want the accountability; the responsibility for resolving chargebacks and disputes; the burden of being supervised by financial services regulators; the expense of staffing a call center. Nor do they have the ability to act on that accountability, because taking actions requires a regulated bank account, a ledger of record, and direct access to payment rails and data that banking institutions have spent years building.
Agentic AI could become a dominant interface layer in financial services. That much seems highly likely. In the future, we will have primary banking agents.
What’s less settled is which companies consumers will choose to provide those agents.
Will we choose a centralized/decentralized structure, where an AI agent orchestrates across a myriad of constantly shifting back-end accounts the consumer has no visibility into?
Or a centralized/centralized structure, where the same, trusted company provides both the primary bank account and the primary bank agent, in a fully aligned and transparent way?
I’d bet on the latter. The primary bank account question was never really about interest rates or branch density. It was always about trust.
The institutions that earn the trust to provide both the account and the agent will win the next decade.
