Banker in Your Pocket

The distinction

The formulation belongs to David Velez of Nubank and is relayed by Micky Malka as the best available compression of where consumer finance goes next. Mobile put a bank in your pocket: your balance, your card controls, your transfers, available anywhere, which was mostly a distribution and interface achievement. AI puts a banker in your pocket: judgment, someone whose job is to look at a specific situation and say what to do. That kind of service existed before and was rationed by wealth. The claim is that AI removes the rationing.¹

What arrives soon

Malka's near-term expectation, stated as something he thinks arrives "very, very soon": "An AI agent that can go and do all kinds of investment decisions, recommendations. He can go do all kinds of commerce decisions and recommendations. And he can start to do any kind of payment decisions and recommendations on your behalf."¹ The doubling in that line, decisions and recommendations, describes a spectrum from advice to execution, with agents specialized by function rather than one general assistant handling everything.

The stated payoff is not better products but "time and ideas that you don't have today," returning to an ordinary customer what a wealth-gated advisor relationship used to cost.¹

Why it matters

It names the specific function being unbundled: not a bank's balance sheet and not its interface, but its advice function, historically the highest-margin, lowest-scale, and most wealth-gated part of the business.¹ If AI commoditizes that function, the competitive question shifts to who owns the relationship in which advice is given and acted on, a different fight from the one fintech has been having about interfaces and interchange.

The idea also raises questions it does not answer. A banker in your pocket implies a conversation, yet an agent making investment, commerce, and payment decisions on someone's behalf is exactly the kind of activity that raises suitability, fiduciary, and recourse questions the formulation does not touch. And the line itself is a slogan more than a forecast: it supplies no evidence about timing, and "very, very soon" carries different weight coming from an investor whose own fund would back the companies that build it. The claim also arrives secondhand: Velez's words relayed by Malka, with no Nubank product yet named as evidence.