Want unlimited access to top ideas and insights?
Jon Lofthouse will tell you he's only ever had two real jobs.
For nearly three decades he worked at Citi, where he launched his banking career at Salomon Brothers before it was folded into the firm in 1998. Before that, he spent a year building software used to refuel nuclear power stations.
"I've had an interest in technology for a long time," he said. "My parents bought me a computer when I was about seven—a ZX Spectrum." It was one of the most popular home computers in the U.K.throughout the 1980s, sort of the British equivalent of the Commodore 64 in the U.S.
At Citi, the former CIO has had an outsized impact on the technology powering America's third-largest bank, with $2.67 trillion in total assets. [Editor's note: After his selection as one of the Most Innovative People in Finance, Lofthouse left Citi.] His mission was to simplify the bank's core technology systems. In 2025 alone, he helped retire or replace 384 different applications to streamline everything from risk analytics to payments infrastructure and client services.
Early in his banking career he helped pioneer the first generation of fixed income electronic trading. "When I first joined the firm, trading in bonds was something you did on the phone—or perhaps a fax machine," he recalled.
Next came the 2008 global financial crisis. "Lehman happened, which made the world very interested in counterparty credit risk," he said. "We realized we needed a risk platform that spanned all of our markets businesses," from commodities to foreign exchange. So he and his team built one.
The platform, called XiP (pronounced 'zip'), functions as a centralized analytics and risk engine, unifying risk calculation across asset classes rather than processing them in siloed systems. Citi says the platform could potentially process billions of calculations a day.
Since 2024, Lofthouse led the effort to weave AI into every layer of Citi's business, rolling it out to 182,000 colleagues in 84 countries. He said that automating routine tasks gives employees more time to focus on higher-value, more creative work.
So far AI has had some of the biggest impact on how the company codes. Citi estimates that these tools have facilitated 1.5 million automated code reviews, which has freed up 100,000 labor hours a week. Lofthouse says this allows developers to focus more on innovation.
AI is also driving faster and more effective client interactions, Lofthouse added, shaving minutes off of each exchange. "2025 was the year we moved the needle on all those metrics," Lofthouse said. "People got to where they needed to be, and got actions taken faster than with our previous solutions."
Finally, Citi built an AI platform called Citi Stylus Workspaces that analyzes and summarizes documents for content generation, research and brainstorming.
Citi may be all-in on AI, but it's not interested in placing any bets on which of the underlying technology will prevail.
The bank built a model-agnostic layer that lets it swap between different AI platforms when it wants to, without having to recode everything.
"The space is moving fast," he said. "When the next best model shows up tomorrow, we can just plug it in and make it available across all of our use cases."
