- Citigroup slashes document review from over an hour to 15 minutes with AI automation.
- AI automates coding, testing and legacy data migration across 50 internal processes.
What happened
AI reduces onboarding time and accelerates replacement of legacy banking systems
Citigroup is deploying artificial intelligence to accelerate account openings and modernise its technology stack, according to its head of technology, Tim Ryan.
AI tools now automate document processing and onboarding workflows. In one key use case, document review time before opening an account has dropped from more than an hour to just 15 minutes in its U.S. services division.
The bank is also using AI to migrate data from legacy systems, automate coding tasks and speed up software testing. This supports a broader effort to retire outdated infrastructure and improve productivity.
Citigroup has identified around 50 internal processes for automation, including client onboarding and “know your customer” (KYC) compliance workflows.
The shift is backed by major organisational changes. The bank has expanded its technology workforce to about 50,000 employees and is reducing reliance on external contractors from roughly 50% to a target of 20%, with the transition already halfway complete.
These investments also respond to regulatory pressure. Citigroup remains under 2020 consent orders requiring stronger risk controls and improved data governance, overseen by the Federal Reserve and Office of the Comptroller of the Currency (OCC).
Why it’s important
AI becoming embedded in core banking operations
Citigroup’s rollout shows AI is no longer experimental. It now underpins core banking workflows.
The combination of onboarding automation and system modernisation marks a shift towards fully integrated operational AI. This reduces manual processing, shortens service cycles and improves scalability.
Embedding AI into legacy replacement also changes how banks upgrade infrastructure. Automation now accelerates migration, testing and compliance simultaneously.
This model strengthens both efficiency and regulatory resilience. It signals a broader industry move towards AI-driven operating models, where productivity gains come from end-to-end workflow automation rather than isolated tools.
Also read: Uber taps Amazon’s custom AI chips for ML workloads
Also read: OpenAI acquires tech talk show TBPN in unexpected media push
