JPMorganChase expands Claude Code as strict production gates remain

By Nathan Brooks

Anthropic announced on November 9, 2026, that JPMorganChase is rolling out Claude to employees beyond engineering, with Claude Code at the center. The expansion extends AI-assisted work into product development, infrastructure, risk assessments, and Asset & Wealth Management. According to Anthropic, broader access comes with strict limits on what agents can reach and who may approve production changes.

Claude Code moves into core workflows

Anthropic places Claude Code at the center of the engineering rollout. The company describes the coding agent as one that plans work, writes code, runs tests and opens pull requests. In the Corporate and Investment Bank, Anthropic says work is moving into production faster, with added capacity redeployed against business priorities.

Engineers are modernizing additional legacy systems and sharing their own agents, Anthropic said. The shared tools include pull-request review and self-healing.

Security is part of the deployment

Anthropic says JPMorganChase’s security team uses Anthropic’s frontier models to find and fix vulnerabilities faster. The company points to Project Glasswing as the earlier description of that work. Glasswing was announced on April 7, 2026, with JPMorganChase among launch partners set to use Claude Mythos Preview for defensive security.

JPMorganChase Chief Information Security Officer Pat Opet described an early opportunity to assess next-generation AI tools for defensive cybersecurity across critical infrastructure. Opet said: “We will take a rigorous, independent approach to determining how to proceed and where we can help.”

Work moves beyond the codebase

Anthropic described the expanded use across functions outside core engineering. The company identified the following workflows:

• Infrastructure and operations: Claude Code handles monitoring, alerting and root-cause analysis.
• Product: Teams use Claude to turn requirements directly into engineering backlogs.
• Risk and program work: Teams apply Claude Code to risk-and-controls assessments and program planning.
• Research and planning: Research, requirements, design and planning fall within the expansion.

These uses extend the rollout from code production into operational and planning work, Anthropic said. They do not expand agents’ authority over production approval, which remains human-controlled.

Wealth teams pair business and technology

In Asset & Wealth Management, JPMorganChase is placing employees with business insight alongside technology and product specialists in small, empowered teams, Anthropic said. Claude Code is now part of their engineering workflow.

Anthropic says the structure helps teams deliver customized client investment solutions at scale and move faster. The company says every change faces more rigorous testing before it reaches clients.

A tax connector for advisors

Separately, J.P. Morgan Asset Management announced a connector in October 2026. It brings 55ip’s ActiveTax Technology into Claude for Financial Advisors and lets advisors evaluate the tax impact of moving a client’s holdings into a new portfolio.

The advisor product combines connectors and workflow skills for research, preparation and documentation, J.P. Morgan Asset Management said. Investment recommendations, client communications, compliance determinations and other regulated activities remain subject to human review and approval.

Production stays under human control

Anthropic says Claude Code runs inside Devspace, JPMorganChase’s containerized development platform. The platform runs on the firm’s own infrastructure rather than developer desktops, and agents receive no direct access to employee credentials or internal systems.

Access expands in stages, Anthropic said. Each new Claude model must pass the bank’s approval process before employees can use it.

Thousands of automated tests and approval gates check every change, Anthropic said. An enforced pipeline gate returns a hard pass or halt, so Claude cannot sign off on its own work. The company says people remain accountable for every production change, keeping model output separate from production approval.

What changes for practitioners?

Anthropic said JPMorganChase’s requirements are shaping what it builds for other regulated institutions. The rollout reaches software engineering, backlogs, monitoring, risk work, investment customization and tax analysis, but it does not grant agents autonomous production or client authority.

Practitioners gain agent assistance across more of the workflow, from research and planning through testing and documentation. They still face model approval, automated checks, a pass-or-halt pipeline and human accountability for every production change.

The next test is whether this controlled expansion can retain the faster production pace Anthropic attributes to the Corporate and Investment Bank. As adoption widens, JPMorganChase’s deployment will show whether banks can give agents broader reach without giving up final control over code, advice and regulated decisions.