Jenius Bank

Nobody brags about their pipes. You do not invite people over to admire the infrastructure that makes your water clean or your toilet flush. Plumbing is invisible by design, noticed only in catastrophe. Data is the same. Nobody is in a boardroom celebrating the data dictionary they finally built or the governance framework they got the business to own. But the moment something breaks, the moment the AI model produces nonsense because the underlying data was never properly defined, or the moment the regulator asks a question you cannot answer cleanly, suddenly everyone wants to talk about the pipes.

I have spent over twenty years in financial services, data and analytics. I did not plan that. I started on the phone at a call center, the only person on the floor who knew how to use Excel and PowerPoint, which made me everyone’s favorite free resource. I had no idea I was building a career. I was just being useful.

What I did not know then, and what took me two decades and a greenfield banking startup to fully understand, was that the most important work in data has almost nothing to do with the exciting part.

The Hidden Weight of Data

For most of my career, I was a data consumer, an analyst, then a data science leader, cleaning data, building fraud detection models, launching a cyber threat operations center with zero network security background, and earning patents on work that started with problems we were obsessed with solving rather than solutions we wanted to force onto them. I spent years telling organizations what good data foundations looked like.

Then I joined Jenius Bank, a digital-native institution within SMBC Group, and for the first time I was responsible for the whole thing. Not just consuming the data, but building the environment it lived in. Eating my own dog food.

 When your data has lineage, when definitions are documented and when the business owns what the data means, you move faster because you stop rebuilding trust after every audit finding. 

What I found was humbling. Every decision I made had downstream consequences I had never considered from the consumer side. And the instinct of the business teams around me was completely understandable but almost completely wrong: load everything. Cover your bases. Figure out what it means later.

Later, it turns out, it is a Sisyphean exercise. You load data without definition, without structure, without the business owning what it represents, and you will spend the rest of your organizational life pushing that boulder up the same hill. Every migration, every new platform, every AI initiative runs straight into the same wall. The boulder does not care what year it is or what the tool is called.

The Skill Nobody Puts in the Job Description

The core challenge in data is not technical. It is that the people who generate and own the data, the product managers, the operations teams, and the business leaders, have never been asked to think about it before it becomes a problem. They want to report after something is built. Asking them to define data requirements before a system exists is genuinely abstract to them. It is not that they deprioritize data; they do not know how to prioritize it. There is no framework in their mental model for what that would even look like.

Which means the most critical skill for an analytics leader is not statistical. It is the ability to make the business feel responsible for its data without making people feel blamed for not knowing. That is the real translation work. It is harder than any model I have ever built, and it never appears in a job description.

Governance and compliance follow the same logic. In financial services, you operate inside a regulatory environment that does not negotiate with your AI roadmap. Leaders who treat governance as a constraint on innovation are often the ones who never built it properly to begin with. When your data has lineage, when definitions are documented, and when the business owns what the data means, you move faster because you stop rebuilding trust after every audit finding. The foundation is not an obstacle. It is the only thing that makes speed sustainable.

Why Data Enthusiasm Isn’t Enough

Here is what nobody in this conversation wants to say out loud: we are watching the same movie again. In 2004, for many professionals, Excel was new software, exotic even. Now it is the floor, the baseline expectation of every business user. We spent years getting an entire workforce to that level of data fluency, and we are about to do it again.

AI and data literacy are the new baselines, and most organizations are approaching them exactly the way late streaming entrants approached Netflix. They saw the product, got excited, started moving fast, and realized years too late that the people winning had been building infrastructure long before anyone was paying attention. They got their lunch eaten.

Data enthusiasm is not data centricity. Enthusiasm is excitement about what AI can produce. Centricity is the discipline of asking, before anything else, where the data comes from, how it is stored, how often you need it, what every element and variable means, and who owns that definition. Organizations that answer those questions first will eat the lunch of everyone who skipped straight to the demo.

For the professionals early in this journey, the ones who want to lead data transformation and feel the urgency of this moment, you are not behind. You are exactly where I was once. I just happened to be the only person on the floor who knew how to use Excel.