Testing a new credit card or banking product used to require months of regulatory research and customer adoption. Now banks are building customers on their behalf.
According to a report by Global Finance, financial institutions are replacing real customers with artificial intelligence-generated stand-ins. Synthetic profiles cost very little and have none of the compliance risks associated with real customer data. Synthetic consumers do more than just compress timelines. This will change the way banks bring products to market.
This implementation is spreading to major institutions on both sides of the Atlantic. US Bank is introducing synthetic audiences to model consumer segments such as high-net-worth households, testing messaging and refining campaigns before launch, Global Finance reported.
JPMorgan Chase produces synthetic financial data to simulate market behavior for risk management and product design. NatWest, Monzo, and Santander use synthetic data ecosystems to train their AI models.
FCA conducts live AI test, but governance questions remain
In the UK, the Financial Conduct Authority (FCA) is moving to incorporate this practice within its regulatory framework. The FCA’s AI Live Testing initiative launched its first cohort in October, including NatWest, Monzo and Santander. The FCA has announced that it has launched a second cohort in April, adding Barclays, Lloyds Banking Group and UBS. Use cases include payments on behalf of others, anti-money laundering detection, and know-your-customer checks. Testing is expected to be completed by the end of 2026 and the evaluation report will be submitted in the first quarter of 2027.
The FCA described the initiative as the first of its kind in the financial sector, the agency reported. Companies are citing AI live testing as a way to overcome so-called “proof-of-concept paralysis,” where regulatory uncertainty stalls AI efforts.
Advertisement: SCROLL TO CONTINUE
However, sandboxes have their limits. Mudit Gupta, EY’s Americas Financial Services Consulting AI Practice Leader, told Global Finance that many banking industry leaders believe that agent AI can transition more quickly if governance is not perceived as a constraint. “But really, it’s the governance that allows these systems to be deployed at scale,” says Gupta.
Gupta added that synthetic data is often treated as inherently safe. it’s not. Sensitive signals can be leaked through inference and linking risks. It also replicates and amplifies historical biases and embeds them behind layers of abstraction, making them difficult to detect, audit and challenge, he said.
Synthetic data proliferates into finance operations and fraud detection
The large scale of deployment makes governance issues urgent. As PYMNTS reported, this technology has already migrated into treasury and financial operations, where predictive models have traditionally relied on data that quickly becomes outdated.
Regulators are unlikely to treat AI outcomes as abstract or low-stakes. According to a report by PYMNTS, fraud at financial institutions accounts for 71% of incidents and losses due to credential theft and account takeover. These are areas where AI is being deployed to determine identity, authorization, and intent in real time. The FCA has announced that it will publish a report on good and bad practices on AI in financial services in the second half of 2026.
For all of our coverage of PYMNTS AI, subscribe to our daily subscription AI Newsletter.
