For decades, every payment, trade, and transfer ultimately passed a human checkpoint someone who could pause, question, or pull the brake.

Autonomous AI agents remove that checkpoint. They can initiate, approve and settle transactions in milliseconds, at a scale and speed no risk team can manually supervise.

Joshua Rosenberg spent 22 years at the New York Fed, leading its second line of defence as Chief Risk Officer through the financial crisis.

A three-time CRO, he's since carried that playbook from regulation into community banking and now into building a stablecoin bank from the ground up.

The financial services industry has lived through multiple waves of transformation — internet banking, mobile, cloud, and now tokenization and agentic AI. Each time, the promise was revolution. Each time, most banks barely moved.

But something feels different now.

Stablecoins are no longer a crypto story. They are fast becoming the settlement infrastructure that the global financial system cannot avoid — and the policy frameworks designed to govern them are not keeping pace.

Today we are asking the question that defines the next chapter of global banking:

Will the stablecoin transformation actually land or are we watching the same story repeat itself one more time?


Today I am joined by someone who has spent 22 years inside the engine room of global banking and has seen every promise of transformation up close.

Namit Khanna is Executive Director of Strategic Accounts at Thought Machine — the cloud-native core banking platform trusted by some of the world's most innovative financial institutions, Fedwire-certified by the US Federal Reserve, and now part of the Mastercard Crypto Partner Programme.

Namit began his career as a software engineer in the early 2000s, building front-end interfaces on top of core banking systems. He went on to work with Finastra, Oracle, DXC, Finacle, and SAS — helping financial institutions across Asia, Europe, and now the United States modernise their technology stacks and transform how they use data.