The Bank of New York Building cornerstone at 1 Wall Street, Manhattan.
getty
BNY is one of those institutions that underpins the financial system even if most people never see its brand on a street corner. Founded in 1784 by Alexander Hamilton, the firm has spent 241 years as a centerpiece of global markets. Today, it oversees more than $57 trillion in assets across custody and related services, moving money, managing money and keeping it safe for many of the world’s largest institutions.
Leigh-Ann Russell is BNY’s chief information officer and global head of engineering. “We provide the infrastructure on which global markets operate,” she noted with pride. “We are not a retail bank. We are not an investment bank. We move money, we manage money and we keep money safe.”
Russell’s remit spans the full breadth of technology, from cybersecurity and resiliency to business facing platforms and the foundational engineering capabilities that support them. “I cover all of technology at the bank,” she said. “We look after cybersecurity, infrastructure, the resiliency on which the bank runs and all of our business facing technology, plus functions such as AI, data, software, engineering, architecture and controls.”
A Platform Operating Model for a Platform Company
In many large enterprises, technology mirrors business unit boundaries. Russell described a deliberate move away from that structure at BNY. “BNY previously was operated with three big businesses that tended to operate in a bit of a silo,” she said. Under CEO Robin Vince, the company has reorganized around a platform operating model, aligning teams to the client and engineering platforms that power day-to-day operations.
BNY CIO and Global Head of Engineering Leigh-Ann Russell
BNY
A platform can be client facing, such as treasury services or payments. It can also be an engineering platform, including the API layer, architecture layer and software frameworks that standardize how teams build. “We are organized by practice and platform,” Russell clarified. “That’s the classic matrix in a platform model, and engineering itself is a practice.”
That structure informs how she organizes leadership. Client-facing CIOs focus on business operations, while horizontal functions run the engineering practices and guardrails. Russell described the core functions as including “cyber, production services, infrastructure, data, AI, architecture and controls.” The model aims to reduce duplication, increase reuse and make the bank faster without compromising safety.
Resiliency and Cybersecurity as Job One
When $57 trillion of client assets flow through your systems, the margin for error is effectively zero. Russell was unequivocal about the priorities that come with that responsibility. “Our number one job in engineering is to keep the bank safe,” she said. “Resiliency and cybersecurity are job one for all of us in engineering.”
BNY has invested heavily in its resiliency framework and data center upgrades across a hybrid approach that includes both cloud and its own facilities. Russell pointed to measurable improvement from those investments. “We have seen a 90% decrease in our serious incidents; our P1s and P2s,” she noted, crediting forward looking work in automation, self-healing and the use of machine learning and AI for anomaly detection and system health.
The urgency is only increasing. “The threat landscape is only continuing to degrade,” Russell emphasized. Her focus for the next phase is expanding automation, self-healing and AI enabled operations so resiliency keeps pace with the evolving risk environment. In her view, clients will not tolerate delays, instability or uncertainty, especially when the stakes are the stability of global markets.
AI Enablement at Enterprise Scale
If resiliency is job one, Russell’s next priority is enablement at scale. She described BNY’s mantra in plain terms. “AI in the hands of everyone for everything, everywhere is the goal,” she said. The bank set an internal goal of reaching 65% workforce adoption of Eliza, its agentic platform named for Alexander Hamilton’s wife, Eliza. It exceeded that ambition dramatically. “We achieved 99% around midyear,” Russell explained, meaning nearly everyone has received initial training on using Eliza safely and productively.
With broad adoption established, BNY is now going deeper by persona, including 40-hour boot camps. Nearly 1,000 of BNY’s roughly 50,000 employees have completed the intensive training. Russell sees this as a practical path to turning AI from novelty into performance. “I don’t think AI should be an engineering accountability,” she said. “Those who are deep experts in their jobs will figure out how AI can help them by removing mundane activities’ cognitive load.”
That same logic extends into software engineering and leadership expectations. BNY rolled out GitHub Copilot broadly, alongside Microsoft Copilot 365 and has pushed hard to achieve hands-on fluency. “Every single one of our coders, nearabout, is using AI assisted coding,” Russell emphasized. More recently, the bank rolled out Windsurf, a vibe coding tool, to all 7,000 developers. Russell herself took a vibe coding course and brought Windsurf in to run a hackathon with her leadership team. “To be an engineer and leader in 2025 and beyond, you need to go back to being hands-on with the tools,” she recommended.
Leading by example is part of the operating model too. Russell described how CEO Robin Vince and the executive committee went through hands-on training as well. “You can’t ask people to do things that you as leaders don’t do yourselves,” she said. The payoff, in her view, is visible in both speed and morale. “Sheer developer joy and speed at which people are pushing ideas into production is absolutely game changing,” she added.
From Agents to Digital Employees
As adoption grows, the bank is moving from AI usage to AI creation. Russell shared that nearly half the company has built their own AI agents, spanning both technical and non-technical teams. The bank even taught agent building live during a company town hall. “We taught the organization to build an agent on the fly,” she said. “It’s not complex once you’ve had a go at it.” The objective is not experimentation for its own sake. It is productivity, focus and new ways of working.
Russell also described a more ambitious evolution: treating certain agentic capabilities as digital employees. BNY has roughly 120 enterprise grade AI solutions in production and about 120 digital employees, including digital engineers. These digital employees have their own credentials and virtual environments, communicate through email and Teams and report to human managers. “We onboard them with strict permissions,” she said, emphasizing the reality of operating within a highly regulated environment.
One example is vulnerability management, a task few engineers enjoy but all engineering organizations must execute relentlessly. Digital engineers can scan code, identify vulnerabilities and either escalate complex issues to a human manager or fix simple ones and submit a merge request. Russell noted that these digital engineers have already written more than 50,000 lines of code. They are performance managed too. “We retire them if they’re not up to scratch,” she said, framing it as a practical necessity rather than a novelty. The aim is to free people to focus on higher value work, the work engineers actually want to do.
Looking ahead, Russell is watching quantum, both for cryptography and for compute, given the massive processing demands of global custody and settlement. She is also monitoring digital assets and blockchain as payment systems evolve. Still, she is clear about the central thread. “As proud as I am of what we’ve done in ‘25, we feel like we’re just getting started,” she noted. “AI is going to continue to be a big focus for us in ‘26.”
Russell’s own experimentation with AI underscores her optimism. She has built a digital twin using OpenAI’s GPT by journaling daily prompts, feeding in data such as Whoop metrics and using the system as a personalized performance and wellness coach.
If BNY’s story is any indicator, the path forward for AI in regulated industries will be defined by the same combination Russell has prioritized since day one: safety, enablement and builders who are willing to stay close to the tools.
Peter High is President of Metis Strategy, a business and IT advisory firm. He has written three bestselling books, including his latest Getting to Nimble. He also moderates the Technovation podcast series and speaks at conferences around the world. Follow him on X @PeterAHigh.
