BNY Mellon’s Eliza 2.0 Platform Puts Generative AI at the Core of Financial Operations

"Generative AI is going to be the operating system of technology going forward." This foundational assertion, delivered by Sarthak Pattanaik, BNY Mellon’s Chief AI & Data Officer, encapsulates the firm’s aggressive strategy toward enterprise-wide AI adoption. The global financial services giant, trusted with trillions in assets, has moved past pilot programs and integrated large language models (LLMs) deeply into its operational infrastructure via its proprietary platform, Eliza 2.0, built in partnership with OpenAI. The video commentary featuring Pattanaik, alongside Watt Wanapha, Deputy General Counsel & Chief Technology Counsel, Ed Fandrey, Global Head of Sales, and Michelle O’Reilly, Global Head of Talent, details not only a massive technological deployment but a fundamental shift in institutional culture centered on responsible AI and efficiency.

The context of this deployment is crucial. BNY Mellon is not a nimble startup; it is the world’s largest custodian, safeguarding $55.8 trillion in assets. Operating within this highly regulated environment means that the adoption of cutting-edge, often unpredictable, generative AI technology requires a robust, secure, and highly governed internal framework. The firm’s journey, as described by its leadership, began with the recognition that data quality and governance are prerequisites for unlocking data-driven decisions. This focus on foundational integrity ensures that the powerful capabilities of models like GPT-4 are deployed responsibly.

Eliza 2.0 is positioned as the innovation accelerator platform, designed specifically to bring advanced AI capabilities to every employee, regardless of their technical proficiency. The core mantra driving the rollout is "AI for everyone, everywhere," ensuring that the 50,000-plus workforce is equipped with tools that eliminate friction and boost productivity. This strategy moves AI out of the domain of specialized data science teams and into the hands of operational staff, legal counsel, and sales teams.

The integration of OpenAI models, customized and governed through the Eliza platform, provides immediate, tangible benefits. For client-facing roles, the efficiency gains are not merely measured in reduced administrative hours but in the ability to enhance client relationships. Ed Fandrey noted that the AI assistance "allows us to use that time to spend valuable time with clients," enabling higher-quality, strategic conversations rather than being bogged down in data aggregation and manual reporting.

BNY Mellon operates on a staggering scale, serving as the world’s largest custodian, trusted with $55.8 trillion in assets. This unprecedented fiduciary responsibility mandates that their AI deployment strategy prioritize trust and governance above all else.

This commitment to trust is buttressed by an aggressive training initiative. The firm emphasized that responsible adoption is paramount, especially in a sector where data integrity and compliance are non-negotiable. Pattanaik revealed the depth of this commitment, stating that "98% of our company have gone through the training of responsible AI." This near-universal training ensures that employees understand the ethical boundaries, data privacy requirements, and compliance standards necessary when interacting with generative models, transforming potential risks into managed opportunities.

The platform is designed to allow employees to build their own "agents" or customized AI tools tailored to specific departmental needs. Michelle O’Reilly, Global Head of Talent, described her team’s creation of an agent that can rebrand and correct all learning collateral, illustrating how non-technical departments are leveraging the platform for massive efficiency gains. This approach of decentralizing agent creation, while maintaining centralized governance through Eliza 2.0, multiplies the impact across the enterprise, creating use cases in areas ranging from legal compliance and risk management to sales support and internal communications.

The partnership with OpenAI is framed as a strategic necessity for maintaining technological velocity. In a rapidly evolving field, relying on an internal stack alone risks obsolescence. By leveraging OpenAI’s frontier research, BNY Mellon ensures that Eliza 2.0 remains at the cutting edge. Watt Wanapha, the Deputy General Counsel, highlighted the importance of this relationship in being "able to leverage the latest but also getting great advice" on the evolving risks and best practices associated with large language models. This synergistic relationship allows BNY Mellon to focus its internal resources on optimizing the governance layer and customizing agents for financial services, rather than attempting to compete in core model development.

Ultimately, BNY Mellon’s deployment of Eliza 2.0, powered by OpenAI, serves as a crucial case study for how established, regulated enterprises can successfully navigate the complexities of generative AI adoption at scale. By prioritizing a governed platform, mandating comprehensive responsible AI training, and strategically focusing on efficiency gains that redirect human effort toward high-value client interaction, the firm is not merely modernizing its technology stack; it is redefining its operational backbone for the next decade of finance.