The Digital Banker spoke with Julien Crowe, Senior Director of Artificial Intelligence at National Bank of Canada, on why the Bank treats AI as a powerful predictive tool rather than a substitute for strategic decision making — and how disciplined execution, academic collaboration and clear accountability have enabled it to industrialize AI responsibly.
Artificial intelligence (AI) has moved rapidly from experimentation to deployment in banking. Yet for many institutions, the harder work begins after the pilot phase: embedding models into live operations, aligning them with risk frameworks, and ensuring accountability remains firmly human. The question is no longer whether AI can contribute, but whether banks can deploy it at scale without blurring responsibility.
At National Bank of Canada, that distinction is deliberate. Under the leadership of Julien Crowe, Senior Director of Artificial Intelligence, the bank has built what it calls an “AI Factory” — a structure designed to move AI from research into production while keeping governance, ethics and decision-making tightly integrated. The aim is not to replicate human intelligence, but to deploy AI for what it does best: prediction, pattern recognition, and augmentation.
That philosophy underpins the bank’s approach to academic collaboration, risk management, and generative AI deployment. It also reflects a broader Canadian emphasis on trustworthy AI, where performance and responsibility are treated as inseparable.
From academic insight to production discipline
For National Bank of Canada, AI was never intended to remain confined to research labs or proof-of-concept pilots. Crowe describes an approach that positions research and development as an engine for execution rather than abstraction.
“At National Bank, research and development is strategically positioned as an accelerator for AI deployment,” he says. “By aligning R&D projects with business priorities and anticipating future needs, the Bank ensures that innovations developed in collaboration with academia become practical solutions that drive business value.”
That alignment matters in a large institution, where commercial urgency can clash with academic timelines. Crowe argues that the tension is resolved not by diluting research depth, but by directing it towards operational needs. “This approach allows the organization to solve complex problems ahead of time and seamlessly integrate new capabilities into operational teams, bridging the gap between academic ambitions and commercial execution.”
The model draws on Canada’s strength in applied AI research, particularly in Montréal, while recognizing that value only materializes when models are embedded into systems, workflows and decision-making.
Breaking down silos in a regulated environment
As Senior Director of the Bank’s “AI Factory”, Crowe has overseen the challenge of industrialising AI within a highly regulated financial environment. The greatest shift was cultural, he says, ensuring teams worked together with a shared conviction that AI must always serve people — our clients, our employees and our communities.
“For the Bank, using AI technology is both an opportunity and a responsibility,” he explains. “It is an opportunity that this technology can be a powerful driver in creating a positive client and employee experience and enhancing our risk management capabilities. In addition, it is our duty to roll out these technologies responsibly and in line with our organizational values.”
Historically, teams responsible for AI risk oversight and those focused on AI-driven value creation worked in parallel. Industrialization required those functions to converge. “The success of AI industrialization has come from fostering collaboration between teams that would have traditionally worked separately on AI risk management and on AI value creation, and uniting them around a shared mission,” Crowe says.
That integration, he adds, directly addresses client concerns. “By achieving this balance, we address our clients’ concerns regarding the safety of their assets and the confidentiality of their data, while also delivering high-quality experiences that exceed their expectations.”
Ethics, accountability and lessons from healthcare
Crowe’s perspective on responsible AI has been shaped by his academic background in operations research and statistics within the healthcare sector, an experience that continues to inform his work in financial services.
“Experience in healthcare has underscored the importance of confidentiality, fairness, and accountability — values that translate as well to financial services,” he says. “Whether serving patients or clients, the ultimate goal remains a positive human impact.”
That framing places ethics at the centre of AI development rather than as a compliance overlay. “This perspective guides the development of AI solutions at National Bank, ensuring that ethical considerations and responsible practices are at the forefront.”
It is an approach that resonates as Banks face increasing scrutiny not only over what AI can do, but how it is deployed — and who ultimately remains accountable.
Canada’s role in trustworthy AI
Beyond the Bank, Crowe serves on the IVADO board of director, a Canadian research and innovation consortium, giving him a broader view of Canada’s position in global AI development. He believes the country, and Québec in particular, plays a distinctive role.
“Canada — and Québec, with Montréal in particular — play a pioneering role in shaping the next wave of trustworthy and scalable AI,” he says. “Montréal stands out as a major research hub, not only for technological innovation but also for its longstanding commitment to the trustworthiness of AI solutions.”
That emphasis on trust has deep roots. “Initiatives like the Montreal Declaration for Responsible AI, dating back to 2018, reflect a holistic concern for both performance and ethical standards,” Crowe notes. For Canadian financial institutions, this alignment is cultural as much as strategic. “For organizations and financial institutions in Canada, prioritizing trust is part of our culture, as it aligns with client expectations and industry values. At National Bank, trust is not only a regulatory requirement — it is a human commitment to clients, employees, and communities.”
Moving beyond myths — and keeping humans in the loop
One persistent challenge, Crowe argues, is how AI itself is understood.
“The most persistent misconception is that AI possesses human-like intelligence, when in reality it is a powerful predictive technology,” he says. “This misunderstanding fuels intense emotions, from fear to excitement, and can distract from the practical benefits AI brings to businesses.”
At National Bank, realism shapes how generative AI is deployed. “Innovation with generative AI is carefully managed by ensuring that human oversight remains central to all decision-making,” Crowe explains. “While AI enhances our ability to process information and generate recommendations rapidly, people, with their empathy, understanding of context and relational intelligence, always make the final call.”
Crowe sees AI execution and governance evolving through deeper integration across disciplines. “Our focus will remain on seamless collaboration between risk management, value creation, and technical experts to deliver solutions to our clients’ needs.”
The ambition extends beyond financial performance. “Beyond financial metrics, we are deeply committed to strengthening financial literacy and, AI literacy — to empower clients and employees to make informed decisions and navigate a rapidly changing economy with confidence,” he says, “thereby maximizing positive impact for all stakeholders.”
