BNY Mellon announced Monday (Dec. 8) that the bank will integrate Google Cloud’s Gemini Enterprise into its enterprise artificial intelligence (AI) platform Eliza, further expanding the bank’s use of agentic AI to accelerate research, automate complex tasks and improve data processing.
The company said the upgrade will help employees analyze large volumes of information more efficiently and deliver insights faster across business functions.
Sarthak Pattanaik, BNY Mellon’s Chief Data and AI Officer, said the initiative reflects the firm’s broader strategy to embed AI more deeply across the organization. “At BNY, our AI strategy is clear: AI for everyone, everywhere and everything,” he said, describing Gemini Enterprise as a way to “deepen the platform’s agentic research, integrate data sources more seamlessly and securely and elevate the user experience.”
The announcement also said the integration will add multimodal reasoning, stronger security controls and a more intuitive framework for building enterprise-grade AI agents.
Agentic AI Gains Momentum Across Banking
The BNY development arrives as more financial institutions explore AI systems capable of performing multistep tasks under human oversight. PYMNTS reporting shows that banks have begun deploying AI agents internally after early efficiency gains from chatbot automation. Some institutions now use “digital employees” to assist with functions such as payment validation and code remediation, operating through existing governance requirements and audit controls.
Citi’s recent enhancements to its Stylus Workspaces platform follow a similar direction. The bank introduced agentic features designed to help employees complete research and data-driven tasks more quickly by integrating internal datasets, productivity tools and structured reasoning.
Bendigo Bank in Australia deepened its partnered with Google Cloud to introduce Gemini Enterprise across internal workflows and modernize employee tools, citing efficiency gains from early pilots.
BNY Mellon’s integration of Gemini Enterprise into Eliza fits into that pattern. While the bank has previously deployed AI-driven tools for operations and automation, the new capabilities aim to streamline information-heavy processes such as market analysis, internal reporting and synthesis of structured financial content. The upgrade focuses on speed, data consistency and improved access to research materials within controlled, auditable environments.
AI Governance Becomes Central
As banks broaden their use of AI agents, the risk profile also expands. PYMNTS noted that as AI moves from assisting with back-office workflows to supporting processes that reach closer to the balance sheet, such as risk assessments or decisions tied to credit and exposure, institutions must maintain heightened oversight. The analysis warned that errors, biased data or opaque reasoning could create downstream vulnerabilities if governance practices fail to keep pace.
For BNY Mellon, the Eliza expansion sits within a controlled governance framework designed for enterprise use. Google Cloud said Gemini Enterprise offers strengthened security, responsible-AI controls and auditable agent behavior. These features align with the bank’s stated requirement that all AI outputs remain reviewable and transparent, with human supervision embedded throughout workflows.
The broader industry backdrop underscores why that structure may matter. Banks adopting agentic AI stand to gain speed, standardization and efficiency across data-heavy functions. Yet the benefits depend on data integrity, model validation and rigorous oversight. BNY’s description of its Eliza upgrade reflects those priorities, with the bank emphasizing secure data integration and reliable reasoning as critical components of its deployment.
A recent PYMNTS study shows how quickly enterprise adoption is accelerating, with 76% of CFOs reporting measurable productivity gains after pilot deployments of agentic AI. The research found that within 90 days, CFOs moved from exploratory conversations to running real financial workflows through agentic systems, citing a 40% reduction in manual reconciliation work and 30% faster close cycles in early trials.
Nearly three-quarters of executives said the technology improved decision speed, while more than half reported clearer, earlier ROI signals compared to prior automation tools.
Source: https://www.pymnts.com/