Traditional outsourcing models are dying

For decades, banks looked to outsource their operations to low-cost centres around the world. The logic was always cost arbitrage: replace expensive operators in high-cost markets with less expensive operators in emerging markets.

This model was fit for its time. It relied on the assumption that the only way to scale expensive operations was to lean on BPOs for a fraction of the cost. But in an AI world, traditional outsourcing is about to die a quick death.

The reason is partly about cost, but mostly about risk.

Outsourcing is moving to software and AI agents. This will be cheaper than traditional low-cost outsourcing. But more importantly, it gives banks a degree of control, visibility and auditability over their operations that has never existed before.

In a modern digital operating system, operational activities and compliance controls are built into the software itself. A bank built this way can see into every part of its operations in real-time and fix problems at the source. That is a fundamentally different outcome from moving people from one organisation to another and reviewing the failures after the fact.

Constantinople’s banking fabric offers banks this new path. Of course, whenever banks consider a different model, it feels risky at first.

When banks consider moving onto a platform like Constantinople, executives sometimes say: “I feel like I’m giving up control over my bank.”

But the reverse is true.

Most banks today have the illusion of control. They don’t have deep, real-time visibility into their operational activities and compliance controls, and certainly not when those activities are outsourced to low-cost operators. The only option is to layer more people and processes on top of existing people and processes.

Constantinople is different. Our use of software and AI gives banks back control. It gives executives and boards the information and confidence they need to discharge their regulatory obligations. It completely changes the game.

Constantinople’s model also scales. Traditional outsourcing means more hires, more training, more management, more to go wrong. Software scales differently. Once a process is built and a control is codified, it runs for the thousandth customer exactly as it ran for the first, at near-zero marginal cost and without the variability that comes with people and manual controls.

The banks that grew fastest over the past decade understood this. Nubank now has 130 million customers and manages operations for around 10% of the cost of a traditional bank. Revolut has 70 million customers, adding close to a million every three weeks, and manages operations via engineering not offshore staff.

I’ve spent my career in banks, or advising them on tech and how they operate. I joined Constantinople three years ago because this shift is inevitable. The easy response is to reach for the familiar levers. But the familiar levers are now the bigger risk. But I chose to solve the harder problem, leveraging software and AI to help banks solve for scale and risk