Machine learning

TD Bank's new lab seeks to divine the future of banking.

BankMobile, the digital-only subsidiary of Customers Bank in Wyomissing, Pa., is planning to use software by Upstart to offer its first credit product to graduates and other young consumers with little to no credit history.

Key efforts for banks include developing AI applications, adding fintech allies and strengthening cyberdefenses.

For years, American Express has analyzed the more than $1 trillion in annual purchases taking place on its closed-loop network to better advise clients to formulate marketing programs.

Without the constant innovation that artificial intelligence empowers, legacy companies will quickly lose their edge as competitors whisk away consumers with new offerings and customized products, writes Ajay Vij, senior vice president and head of financial services for Infosys.

Neural network systems — which try to emulate the human brain — will become increasingly important as the stolen identities from the recent Equifax breach and others start surfacing in the coming years on fraudulent payment card or loan applications.

As many as a dozen financial institutions are deploying IBM’s Watson to search for signs of employee misconduct so they can avoid a Wells Fargo-size scandal. But the legal and technical limits of its use are major issues.

Banks such as Citigroup, Regions and TD have decided they need to offer mobile customers truly customized experiences. They are experimenting with different ways of doing so that come across as helpful without being intrusive.

Bankers need to set aside their post-crisis aversion to risk, recapture some of the magic they once practiced as innovators, fear fintechs less, and worry about the Googles and Amazons more, says Peter Memon, a bank IT veteran turned global consultant.

Some financial institutions are looking at how to translate the complicated and document-heavy mortgage lending process into the digital world of voice-interactive personal assistants.