Kirsti Racine delved into AI long before it became trending.
Before taking on the role of VP, AI Technology Lead at TD, she completed a master’s thesis in AI at Simon Fraser University and worked at IBM, where her focus was on developing early machine learning systems for healthcare, specifically to expedite patient routing for MRIs.
“I was told AI wouldn’t be a viable career path as it wasn’t popular back then,” Racine shared.
“We’ve dedicated considerable effort to ensure our colleagues feel confident and comfortable with AI,” she added.
Now, Racine leads a team driving AI adoption in one of Canada’s largest banks. The platform she manages, the TD AI Platform (TAP), acts as the core hub for TD’s AI infrastructure.
TAP supports all business areas, including banking, wealth management, insurance, and marketing.
It is designed to make AI accessible to both engineers and everyday staff.
According to Racine, TAP underpins most of TD’s significant AI projects, enabling teams to innovate and streamline tasks more effectively than they anticipated.
“Our finance team used the platform to create use cases that explain financial data in plain English, which were unexpected and delightful,” Racine noted.
Origins of TAP trace back to late 2022, as generative AI was gaining traction globally. TD had previously invested in AI, acquiring Toronto-based AI firm Layer 6 in 2018 and completing a major cloud data platform migration by mid-2024.
“We got a bit lucky at TD,” Racine said. “Incorporating machine learning ops from Layer 6 and leadership from our cloud data platform gave rise to TAP,” she added.
Within months, TD unveiled two early generative AI tools. One is a deep-search assistant that aids contact centre agents in quickly accessing relevant policies and procedures.
Another initiative involved piloting GitHub Copilot, an AI programming assistant, across its developer ecosystem, providing code suggestions and testing assistance. The rollout included comprehensive training, with 92 percent of TD engineers now engaging with Copilot weekly.
“We’ve dedicated considerable effort to ensure our colleagues feel confident and comfortable with AI,” Racine reiterated.
TAP now forms the foundational layer enabling secure and repeatable AI deployment across TD. Despite being a lean team of 60, its impact reaches every part of the organization.
TD runs AI environments tailored to each business line for both predictive and generative AI use cases. Automation has significantly reduced deployment time and costs by up to 85 percent.
One ambitious project is TD AI Prism, a predictive foundation model aimed at enhancing predictive model accuracy and speed. It stands as the most substantial AI model ever deployed by TD, enabled by TAP’s automated production deployment capabilities.
“TD AI Prism sets us apart,” Racine stated. “It’s transformed how we approach predictive AI, allowing unprecedented model deployment at scale.”
TAP was developed by the bank’s internal cloud and data teams and Layer 6’s machine learning professionals, a partnership central to its evolution. Layer 6 provides many models, while Racine’s team focuses on machine learning ops, infrastructure, and frameworks.
Racine and Maksims Volkovs, TD’s Chief AI Scientist and Layer 6 Co-founder, credit their collaboration for enabling business units to own and build AI applications using shared infrastructure.
“Kirsti and I see AI as key to unlocking next-gen solutions for clients, employees, and businesses. We aim to enable AI at scale and at the organization’s edge,” Volkovs said.
Following the launch of the contact centre assistant, teams from all business lines began developing their own generative AI use cases using TAP’s components, which expedited their initiation—a moment Racine found fulfilling.
The 2025 TD AI Insights Report revealed that 64 percent of respondents feel inadequately trained in AI at work, a challenge that shaped TAP’s approach.
Training is embedded in deployment strategy, delivered in formats suitable for each employee, fostering higher adoption and engagement at TD.
The team invests in soft skills crucial for technical transformation, including change management, team collaboration, and user-centered design—treating every deployment as a product undergoing continuous iteration based on real user feedback.
Racine’s team is expanding TAP’s reach with more environments, automation, and new models under development. While current applications focus on internal operations, client-facing generative AI tools are planned for 2026.
Racine envisions AI as an embedded, reliable infrastructure. “I hope TD employees will perceive AI as seamless, intuitive, secure, and supportive in enhancing daily tasks,” Racine concluded.
At TD, we’re developing AI systems that are intuitive, secure, and meet real-world needs. Read more about TD AI Prism.
All photos provided by TD.