Citigroup introduced a new platform called Arc, which will allow the bank to build and scale AI agents across the firm.
These AI agents enhance human judgment by taking on tasks such as research, synthesis, preparation, and execution, reducing manual effort and accelerating how teams operate day to day.
For example, a banker in the wealth management area might spend many hours preparing for a client meeting, gathering portfolio data, analyzing market trends and modeling scenarios. In the future, a team of AI agents will be able to complete this work proactively and deliver the information exactly when it is needed. By eliminating hours of manual effort, the banker will be able to spend more time building deeper relationships with clients and delivering more personalized, timely service. So, the role of the banker evolves more decisively from coordinator to architect and advisor.
In other words, it means focusing AI on the repetitive and the manual, while deepening judgment, sector expertise and client insights.
With Arc, every agent will be monitored, audited, and governed. Citigroup will know what agents are doing, how they are doing it and the value they deliver. Arc will initially be used by developers to build agents for specific use cases to support its businesses and functions. Over time, access to Arc will expand to support specific needs across the firm.
This initiative is part of Citigroup’s goal to make the company the most AI-empowered financial institution in the world.
Along those lines, Citi Wealth unveiled a new product called Citi Sky, an always-on AI-powered member of the Citi Wealth team. The product was built using technology from Google Cloud and Google DeepMind.
Citi Sky is aimed at reshaping how clients access market insights, act on opportunities, and engage with their financial advisors. Over time, it is designed to create a more intuitive, responsive, and personalized wealth experience. It was introduced at the Google Cloud Next 2026 conference.
The product will be made available to Citigold clients starting this summer in a phased roll-out in the U.S. Citi Sky was developed using the Gemini Enterprise Agent Platform and will be integrated over time into Citi Wealth platforms in the U.S. It is built to work alongside financial advisors.
Among key features, CitiSky will offer:
- Financial guidance and support: Delivering timely guidance for clients, including prompts for bank certificate of deposit (CD) maturity events and the provision of market insights from Citi Wealth’s Chief Investment Office.
- Conversational interaction: Enabling natural and dynamic engagement by leveraging Google DeepMind’s latest real-time avatar technology and Gemini’s live audio and video models.
- Multilingual capabilities: Launching in English and Spanish, Citi Sky is designed to scale to additional languages over time.
“We believe Citi Sky will change the model of wealth management,” Andy Sieg, head of Wealth at Citigroup, said. “For decades, managing your financial life meant navigating apps, calls, and meetings. With Citi Sky, you simply ask – and act. This is the shift from interface to intelligence, from transactions to outcomes.”
Citi Sky’s development involved Google’s full AI stack, from Cloud infrastructure to Google DeepMind. This included using Gemini models on Gemini Enterprise Agent Platform. To power the Citi Sky virtual persona and multimodal capabilities, Citi collaborated with Google DeepMind’s Applied team and Google Cloud Forward Deployed AI Engineers.
“The future of financial services lies in the ability to turn vast amounts of data into conversational, actionable intelligence for investors,” Thomas Kurian, CEO of Google Cloud, said. “With Gemini Enterprise as the backbone of Citi Sky, combined with frontier models from Google DeepMind, Citi Wealth is establishing a new blueprint for how agentic AI can drive high-quality, personalized financial insights for millions of its customers.”
The technology builds upon Citi and Google Cloud’s multi-year strategic partnership to support the bank’s digital strategy.