As chief of AI and analytics at TD Bank, Luke Gee is working to identify processes that can be streamlined through AI.

The $1.9 billion bank is deploying AI initially for software development and customer assistance and plans to expand AI use to further improve operations, Gee told Bank Automation News.

Toronto-based TD has also urged employees to pitch ideas on ways AI can be used to make their jobs more efficient, Gee said, adding that the bank has implemented 10,000 of those ideas out of 100,000 submissions in the past five years.

“We have iD8, our colleague ideation program, where we crowdsource ideas from colleagues,” Gee said. “If their idea is selected, our colleagues can track the idea from concept to implementation in the platform.”

In an interview with BAN, Gee shared his plans on how AI can be developed and used within the bank to create value. What follows is an edited version of that conversation.  

BAN: How has TD’s gen AI pilot been faring since its launch in May 2024? 

Luke Gee: Seeing colleagues experience the transformative nature of this technology firsthand and imagining a new normal has been exciting for the bank and colleagues. With these programs, we see developers being able to delegate simple, repetitive tasks to Github Copilot and finding more time to focus on functionality; and front-line colleagues seeing guidance on how to help resolve a complex customer issue appear on the virtual assistant in seconds. 

Early lessons learned include the need for deep cross-functional collaboration to bring use cases to life. Gen AI requires strong, agile interaction structures to synchronize technology, risk, business and design expertise. Gen AI also brings with it new considerations around AI risk and appropriate guardrails, the need for new processes for development and deployment, and heavy investment in infrastructure and compute. 

BAN: What is TD’s outlook on gen AI adoption? 

LG: Gen AI is in high demand and adoption is accelerating in the financial sector. Gen AI has the potential to unlock insights and value from data that in turn can enable banks to deliver new capabilities and experiences.

At TD, we are exploring additional use cases to help drive colleague productivity and increase operational efficiency. Gen AI introduces significant changes to the ways we work, and adoption combined with effective change management is critical to long-term success. 

BAN: How does TD find and retain AI talent? 

LG: The market for talent is highly competitive. While AI literacy has been increasing, the need for more specialized skill sets, like machine learning operations, engineering, research science, is also increasing and are in high demand.  

Successful gen AI delivery requires a multifaceted approach that cuts across many functions — business strategy, product management, engineering, science and human-centered design, to name a few. In addition to raw technical skill, practitioners also need strong communication and collaboration skills.

TD maintains close relationships with AI-focused academic organizations that act as a talent pipeline. Through internship programs and research collaborations with a number of universities, we draw in early career talent and invest in their skill development. 

Gen AI is going to reset the productivity bar for banking talent at TD. Automating monotonous tasks will free up colleagues to focus on higher-value work that fully leverages their creativity. Gen AI is a complement to our people, a tight relationship between team members and AI that together helps create a powerful, augmented customer experience.

BAN: What are the benefits of building AI in-house as opposed to outsourcing? 

LG: Building in-house deepens our internal expertise around gen AI and is a very exciting and challenging area for our scientists to dig into, which has helped with talent development and retention. TD has a strong talent bench that works on developing cutting-edge applied solutions.

There are also inherent advantages to in-house models from a governance perspective. We have stronger visibility into what methods and tools are used, better data privacy and security, and more oversight and control to manage changes to the model.

BAN: How does TD consider AI compliance while developing products? 

LG: Our relationship with our customers is based on trust. We need to introduce gen AI carefully. We can’t risk doing so without ensuring our customers are protected and that we do not lose their confidence.

The regulatory environment around gen AI is still evolving. We work with regulators to keep them informed on how we manage the risks associated with emerging technologies.

Register here for early-bird pricing for Bank Automation Summit 2025, taking place March 3-4 in Nashville, Tenn. View the full event agenda here.