A core banking replacement is the single most expensive, highest-risk thing a bank can do. It makes M&A look simple.

I've watched banks spend hundreds of millions on new core systems and end up with basically the same operational problems they had before. The technology is newer. The processes are just as broken.

The reason is almost always the same: CEOs don't ask the right questions at the beginning. They rely on vendors and consultants to guide them, and vendors and consultants have their own incentives.

Here are the questions I wish every CEO asked before signing.

The questions CEOs don't ask before a Core Banking System replacement

What happens to our data when we migrate?

Every bank says they have a data migration plan. Few of them have a good one. The reality is that legacy systems contain decades of inconsistent, incomplete, often contradictory data. Migrating it to a new platform without cleaning it just moves the mess.

Ask to see the data mapping. Ask what gets left behind. Ask how you'll know the migration worked.

Who actually owns the business requirements?

In most core replacements, the requirements document is written by consultants, reviewed by IT, and signed off by business leads who are too busy to read it carefully.

That's a recipe for building a system that matches the document but doesn't solve the actual business problem.

The business needs to own requirements. Not review them. Own them. If your head of retail banking can't explain in detail what the new system needs to do for her teams, you're not ready.

What's the real cost of customizing vs. adapting?

Vendors will tell you their system is configurable. What they mean is: you can customize it. And customization is where core replacement budgets go to die.

Every customization adds cost, complexity, and future maintenance burden. Before you approve any customization, someone should be able to explain why the standard functionality won't work — and why the customization is worth the long-term cost.

Who's still going to be here in three years?

Core replacements take years. The people who start them are often not the people who finish them. That's fine as long as knowledge transfers. But it rarely does.

Ask about succession planning for the key people on the program. Ask what happens if the program director leaves. Ask how institutional memory gets preserved through turnover.

What does success actually look like?

Not in slide deck terms. In operational terms. Faster loan origination? Fewer manual touches? Lower cost-to-income ratio? Shorter customer onboarding times?

If you can't define success in measurable terms before you start, you won't know whether you achieved it when you finish.

Core replacements are hard enough without walking in blind. Ask these questions early, and you might avoid the worst of it.