Citi's Agentic AI Platform Shows How Large Enterprises Are Controlling AI Agent Use
Citi launched an internal agentic AI platform, showing how enterprises are trying to control AI agent access, scope and governance.
In this briefing
Why it matters
Axios reported that Citi has launched an internal AI platform called Arc, designed to let employees build and deploy AI agents securely across the organisation.
What changed
The bank is using a controlled platform approach rather than letting AI tools spread without oversight. The reported use cases include portfolio data, market trend analysis and scenario testing.
The customer service angle
The same pattern applies to support operations. AI agents need access to business context, but that access must be controlled, logged and limited to the right tasks.
What businesses should watch
For customer service teams, the lesson is to choose AI tools that support permissions, source control and escalation. Uncontrolled agents can create accuracy, privacy and compliance risks.
Sources
More AI support news
Build AI Support on Trusted Sources
DocMind helps teams turn websites, help centres, PDFs and policy pages into grounded AI answers with clean escalation paths.
Start Your Free Trial