BNY Mellon Integrates Google Gemini 3 into AI Platform Eliza

BNY Mellon Expands AI Capabilities with Gemini 3

BNY Mellon is advancing its enterprise AI initiatives by integrating Google Cloud’s Gemini 3 into its internal AI platform, Eliza. This integration aims to empower over 20,000 employees to develop and deploy AI agents that automate routine tasks, enhance research, and improve client outcomes.

Background

The Eliza platform, named after the bank founder’s wife, has been central to BNY Mellon’s AI strategy since 2023. It provides tools, data connectors, and guardrails for employees to safely create AI assistants for tasks like payment validation, code repair, market analysis, and client reporting. Eliza already supports numerous automated tasks and tens of thousands of users. The integration of Gemini 3 is designed to enhance Eliza’s capabilities, allowing agents to handle more complex processes.

What’s New: Gemini 3 + Eliza

• Integration Details: Announced in December 2025, BNY Mellon has incorporated Google Cloud’s Gemini Enterprise, including the Gemini 3 model, into Eliza. Gemini 3 is a state-of-the-art multimodal model with advanced reasoning and coding skills.
• Agentic Capabilities: The integration allows Eliza-built agents to autonomously perform multi-step tasks, such as researching disclosures, synthesizing reports, and generating documents, thereby reducing manual labor.
• Scale and Adoption: Over 20,000 employees are using Eliza, with nearly the entire firm trained in generative and responsible AI, reflecting a broad push to operationalize AI.

Strategic Rationale and Industry Context

Financial institutions are rapidly deploying AI agents due to the data-intensive nature of financial workflows. BNY Mellon’s approach of combining advanced models and partnerships mirrors industry trends, balancing capability, governance, and cost. The Gemini 3 integration enhances Eliza’s task handling and demonstrates BNY’s commitment to leveraging leading AI providers.

Risks, Governance, and Responsible AI

BNY Mellon emphasizes responsible AI training and governance controls within Eliza. However, the move to agentic systems increases operational and compliance risks. Tight logging, human-in-the-loop checkpoints, and robust data-access controls are necessary to prevent errors and ensure compliance.

Implications for Clients, Competitors, and the Market

• For Clients: Faster research and automated tasks could lower costs and improve service turnaround times.
• For Competitors: The integration indicates that large banks will continue to pursue multi-model strategies, potentially accelerating feature rollouts.
• For the Market: Successful deployment of Eliza agents could lead to productivity gains and competitive differentiation for BNY Mellon.

What to Watch Next

1. Adoption Metrics: Monitoring the number of live production agents and their efficiencies.
2. Regulatory Scrutiny: Observing regulatory responses as agentic AI scales.
3. Technical Evolution: Potential expansion to other multimodal models or proprietary developments.
4. Market Impact: Changes in client satisfaction and competitive positioning linked to Eliza.

BNY Mellon’s integration of Gemini 3 into Eliza marks a significant step in operationalizing AI agents across the financial sector, aiming to make “AI for everyone, everywhere” a practical reality while managing governance and risk.