Chase CIO Reveals 3 AI Priorities Beyond Code Writing
How JPMorgan's tech leader is redefining engineering roles with agentic AI.
1Engineering roles redefined
Chase CIO Gill Haus states engineers are no longer hired primarily to write code, but to determine what code needs to be written. This shift reflects AI's ability to handle coding tasks through agentic platforms. The bank's engineers now focus on higher-level architectural decisions while AI handles implementation. This change occurred rapidly, with significant impacts visible within just six months.
2LLM Suite deployment
JPMorgan launched its internal LLM Suite in summer 2024, providing secure access to both frontier and open-source language models. The platform enables employees across departments to query documents, create specifications, and develop new services. Unlike siloed AI tools, this enterprise-wide system integrates with existing workflows to remove operational bottlenecks while maintaining strict security protocols required in banking.
3Customer experience transformation
Chase embeds AI throughout customer interactions, from call routing using intent recognition to personalized digital services. The technology operates invisibly, improving experiences without customers noticing the underlying mechanisms. For instance, machine learning analyzes callers' needs in real-time to connect them with appropriate representatives faster than traditional IVR systems.
4Security as prime directive
As a regulated financial institution, Chase maintains strict guardrails around AI deployment. All data usage requires explicit customer consent and complies with privacy regulations. The bank applies the same rigorous standards to agentic technologies as traditional systems. Security teams continuously update protections and address vulnerabilities proactively, especially in customer-facing applications.
5Probabilistic outputs challenge
Agentic AI's probabilistic nature creates reliability challenges for financial services. Unlike recipe suggestions where variations are tolerable, banking transactions demand consistent accuracy. Chase addresses this through human oversight in critical processes and rigorous outcome validation. The bank maintains human-in-the-loop systems for financial decisions where incorrect outputs could have serious consequences.
6Architectural standards evolution
With AI generating code, Chase emphasizes architectural governance over implementation. Engineers now focus on defining patterns, components, and scalability requirements rather than writing individual functions. This shift demands new skills in evaluating AI-generated outputs and ensuring holistic system design. The bank trains staff across departments to use agentic tools while maintaining architectural integrity.
7Workforce upskilling strategy
Chase trains all employees on AI tools, not just technical staff. Product managers learn to specify requirements in ways AI can implement, while engineers develop skills in validating machine-generated code. This broad upskilling enables the organization to redistribute work optimally between humans and agents. The approach aims to eliminate tedious tasks while elevating human roles to oversight and creative functions.
8Future-proofing through agents
Haus views agentic AI as fundamental to remaining competitive in financial services. By embedding these technologies into operational processes now, Chase positions itself to scale services efficiently as capabilities advance. The bank balances innovation velocity with responsible deployment, particularly regarding data security. This strategy reflects a broader industry shift where AI augments rather than replaces human expertise.
Original article: Latest news — Chase CIO Reveals 3 AI Priorities Beyond Code Writing