Citi's AI Overhaul: From 15-Minute Accounts to $3T Infrastructure
Export as clean Markdown. Drag & drop into ChatGPT, Claude, or Gemini.
Citigroup is using artificial intelligence to simultaneously streamline its own operations and position itself as a premier financier of the AI infrastructure boom. The bank's head of technology, Tim Ryan, disclosed that AI has already slashed document review time for U.S. account openings to 15 minutes, while the firm has formed a dedicated AI infrastructure group to compete for what it estimates could be a $3 trillion buildout by 2030. For investors, these two developments are connected: Citi is betting that the same AI transformation reshaping its own balance sheet will generate the advisory fees and financing mandates that could drive a multi-year revenue cycle.
AI Cuts Citigroup's Account Opening Review to 15 Minutes, Targeting 50 Core Processes
The most tangible catalyst at Citigroup is operational: an AI document-processing system has reduced the time needed to review materials before opening a U.S. account in its services division from over an hour to 15 minutes. Tim Ryan, the bank's head of technology, said the system is part of a broader initiative covering Citi's first 50 critical processes targeted for automation, including client and employee onboarding and "know your customer" compliance checks. This matters financially because onboarding speed directly affects the cost of customer acquisition and regulatory compliance, two of the most expensive variables in banking at scale. Citi operates across nearly 180 countries and moves trillions of dollars daily, so even modest per-transaction time savings compound rapidly across the enterprise. The firm has also shifted its technology workforce composition: contractors made up roughly 50% of its tech staff a year ago, and management is "halfway through" a plan to reduce that ratio to 20%, replacing them with in-house software engineers, with a tech workforce now totaling around 50,000 people.
182,000 Employees Now Use AI, With a Centralized Platform Targeting 30%-Plus Efficiency Gains
Beyond process automation, Citi has scaled enterprise AI broadly. Shobhit Varshney, the firm's Global Head of Artificial Intelligence, who joined Citi in September 2025, said that more than 182,000 employees now have access to AI tools, with over 70% actively using them. The firm's strategy emphasizes outcome-first design rather than pilot-led experimentation: leadership defines the desired business result and the expected value improvement before building any solution. Varshney said the goal is often to target efficiency improvements of 30% or more, adding that the mindset shift required is from "a faster scrubber to a dishwasher," meaning redesigning processes entirely rather than accelerating legacy workflows. To embed that culture, Citi created a network of approximately 4,000 AI "accelerators and champions" embedded across business units, peer-level practitioners who help colleagues adopt tools in daily workflows rather than relying on top-down mandates.
Citi Estimates $3 Trillion AI Infrastructure Buildout by 2030, Forming Dedicated Group to Compete
If the internal AI rollout is the operational catalyst, the external opportunity is even larger. According to an internal memo sent in late February by leaders of Citi's investment banking unit, the bank estimates the AI infrastructure buildout will require $3 trillion by 2030. In response, Citi is establishing a dedicated AI infrastructure group to break through internal silos and evaluate "all pockets of capital" as data center deals grow larger and structurally more complex. The group brings together bankers from investment banking, corporate banking, and financing to compete alongside Morgan Stanley, Goldman Sachs, and JPMorgan, which are already building comparable cross-disciplinary teams. The key milestones to watch are Citi's ability to win lead mandates on large-scale data center and energy financings, and whether its internal AI transformation can visibly reduce cost-to-income ratios in upcoming quarterly results.