Artificial intelligence (AI) is changing the way we work, bank, and live. In 2025, TD recognized an opportunity to leverage AI-led innovation to enhance their client’s experience.
This culminated in the launch of TD AI Prism, a powerful predictive model that employs generative AI tools across various business units.
In this Q+A, we talk with Maksims Volkovs, Senior Vice-President, Chief AI Scientist at TD, and Co-Founder of TD’s AI lab Layer 6 to explore how their research and advancements shaped TD’s AI strategy, the impact of AI on client and colleague experiences, and the bank’s commitment to responsible, trustworthy AI practices as it moves into 2026 with ambitious goals.
Can you give me an overview of the AI innovations that TD rolled out in 2025?
MV: In 2025, TD accelerated its AI innovation strategy with a bank-wide commitment to reimagine the banking experience for clients, colleagues, and communities. We launched TD AI Prism, a foundational predictive model leveraging large-scale AI and bank datasets to better understand client needs and introduced seven Gen AI virtual assistants to colleagues across contact centers, branches, TD Wealth, TD Insurance, and TD Securities.
These advancements streamline workflows and empower teams to deliver faster and more personalized services. Additionally, more than 25,000 colleagues were equipped with Gen AI productivity tools, including GitHub Copilot for engineers and Microsoft Office 365 Copilot across the organization, boosting efficiency, fostering innovation, and transforming how we deliver for our clients.
In what ways did the 2025 pilot programs of TD AI Prism and other AI initiatives reinforce TD’s strategic direction for AI in 2026?
MV: From the outset, our strategy was to invest in foundation predictive models and strong results from the 2025 pilot programs with TD AI Prism reinforced that direction. By centralizing our data on TD’s secure cloud platform, we made it much easier to develop and deploy models at scale, which sped up innovation across the bank.
The generative AI pilots—especially in our contact centres—showed us just how much manual work could be automated, and we saw encouraging results in Personal Banking too. These experiences have only strengthened our commitment to this approach as we move into 2026, with an ambitious goal of delivering $1 billion in value from AI.
What’s clear is that having a robust data foundation, reusable AI capabilities, strong oversight through our Trustworthy AI framework, and a focus on business goals are essential—and that’s the blueprint guiding us this year.
How do you think the widespread deployment of AI solutions in 2026 will transform the client experience?
MV: AI is at the core of how we’re reimagining banking for clients in 2026. Our vision is to make banking simpler, faster, and more intuitive. As AI becomes more widely deployed across the bank, clients can expect shorter wait times and faster resolutions when they bank with us. We’re scaling up our Generative AI capabilities to empower our teams, reduce manual work and enhance our colleague and client experience.
We’re already seeing the benefits. Our Gen AI virtual assistants in our contact centres are helping our colleagues get answers quickly, reducing wait times and minimizing the need for call transfers by 12%. We’re also encouraging more clients to complete routine banking tasks—such as cheque deposits—on mobile devices rather than in-branch and we’re working to shift a portion of the more than two million monthly service calls we receive to our app, further streamlining operations and enhancing client convenience.
With the transition to a predictive foundational model, how has TD improved the scalability and efficiency of its AI-driven predictions?
MV: A core element of our AI strategy is to build once and use many times. Traditional predictive AI models are typically built to perform a single task and so each time we wanted to create a new use case, we’d have to build a new model from the ground up. Foundational models like TD AI Prism solve this problem because they process far more data than legacy models and generalize to many predictive tasks.
TD AI Prism can be finetuned and redeployed again and again. We no longer have to build 10 different models for 10 different use cases. TD AI Prism improves predictive performance by ingesting over 100 times more data to understand client needs and so far, it’s proven to be 20-30% better at identifying them than traditional predictive models.
As AI solutions become more integrated into TD’s operations, how is the Bank ensuring client data privacy and maintaining trust in 2026?
MV: TD is committed to upholding the highest standards of client data privacy and trust. We have multilayered safeguards in place and adhere to industry regulations and privacy laws, supported by a strong governance, risk, and controls framework. Every AI solution undergoes rigorous model reviews and process approvals before implementation, including third-party risk assessments.
What steps has TD taken to ensure responsible AI use, and how does the Trustworthy AI team contribute to maintaining ethical standards in 2026?
MV: Our Trustworthy AI team is at the centre of our efforts to deploy AI in a way that maintains the trust we’ve built with our colleagues, clients and communities. This team is made up of AI scientists who are solely dedicated to this cause.
When we invest in AI, we’re not just thinking about what’s technically possible and how to deploy it. We’re thinking about building systems people can trust. Before we introduce new AI models, they have to clear a rigorous series of checks and balances established by our Trustworthy AI team. They established a governance framework that embeds privacy-enhancing technologies, fairness constraints and anomaly detection directly into every AI solution we deploy. Our scientists evaluate every model for privacy, security, fairness and accountability to keep them aligned to human values. They continue to monitor our models even after they’ve gone to market.
Our Trustworthy AI team will continue to be pivotal to our AI operations in 2026 as we continue to introduce new generative AI solutions from our pipeline and begin our exploration of agentic AI in a way that is transparent, fair and accountable.
How does TD envision its role as a leader in AI innovation within the financial sector in 2026 and beyond?
MV: Our vision is to make banking simpler, faster, and more intuitive. We’re already delivering on that promise, and AI is at the core of how we are reimagining banking for clients and colleagues. We’re not new to this. We have been engaged with AI for some time and with the acquisition of Layer 6 in 2018, we’ve built a world-class AI research centre, with our scientists earning recognition at top forums and winning international AI challenges.
We recently opened a New York office to attract top talent and grow our AI team. This commitment is fueling innovation at TD and plays a big role in our success in launching seven Gen AI virtual assistants across multiple lines of business last year. TD was also ranked the top AI patent filer among Canadian banks by Evident AI, underscoring our leadership in AI innovation.
We know AI is not just shaping the future of banking. It’s delivering real, measurable impact today. The banks that move fastest, scale responsibly, and focus on end-to-end transformation will define the next era of global financial services. Our goal is to be on the cutting-edge of this evolution.
Leave a Reply