I spent 25+ years building technology organizations inside existing constraints. Here is what I would do without them.
Not starting with the technology. Starting with the organization.
Most transformation programs fail because org design is treated as a downstream problem. You cannot build a modern capability on a model designed for a different era.
Here is what I would do differently.
HIRE THE SKEPTICS FIRST.
The engineers who raise architectural concerns before a go-live are not blockers. They are your early warning system. Build the culture before you build the platform.
BUILD THE DATA ORGANIZATION BEFORE THE APPLICATION ORGANIZATION.
In intelligence work, the collection architecture determines everything downstream. Banking is no different. Most institutions have it backwards — applications first, data strategy as an afterthought. That is why AI transformation stalls. You cannot build intelligence on a foundation you do not own.
MAKE THE CIO/CTO AND CHIEF RISK OFFICER CO-OWN ARCHITECTURE DECISIONS.
Technology risk is business risk. Treating it as an IT governance problem is how you accumulate 400-application portfolios that nobody fully owns — and nobody can retire without a multi-year program.
BUILD POLITICAL NAVIGATION INTO LEADERSHIP DEVELOPMENT FROM DAY ONE.
Technical excellence without institutional fluency is incomplete preparation. I watched people I had sponsored struggle in senior roles I thought they were ready for. The gap was never the craft.
MAKE “I DON’T KNOW YET” A PROTECTED PHRASE.
Executives who cannot say it are not strong. They are expensive. Build a culture where uncertainty surfaces early — not managed privately until it becomes a crisis.
The technology choices — cloud architecture, AI platform, payment rails — are consequential but tractable.
The organizational choices are where the real leverage is. And where most banks are still getting it wrong.
What would you add?