TD Bank (TD-T) has released its first agentic artificial intelligence (AI) model, geared toward streamlining the application process for its lending services.

The AI model automates the pre-adjudication process for mortgage and home equity line of credit applications. It generates application summary memos for underwriters by performing tasks such as classifying client documents, drawing critical information, calculating client income and searching for discrepancies.

Early results suggest the agentic AI is slashing the workload of TD underwriters from an average of 15 hours to less than three minutes, the Toronto-based bank said in a Thursday morning announcement, while improving accuracy. This, the bank said, reduces risk and the unit cost on adjudication.

Layer 6, TD’s AI research and development centre based in the MaRS Discovery District in Toronto, developed the model in partnership with its data, risk management, global technology and solutions, and real estate and secured lending teams.

"By providing confident decisions earlier in the homebuying process, one of life's milestone moments, we're meeting our clients where they need us most, making the experience simpler and faster,” Mohit Veoli, senior vice-president of real estate secured lending at TD, said in the announcement. 

Canadian banks rushing to AI adoption

Like its banking peers and the corporate world at large, TD is embarking on a widespread adoption of AI in its operations, driven by the promise of increased efficiency and productivity gains. BMO, for example, released Lumi Assistant in 2025, an AI-powered tool designed to simplify access to policy and procedural information across its Canadian personal and business banking.

To this end, TD equipped its customer service agents and software developers with generative AI, and piloted machine learning to pre-approve mortgages and home equity lines of credit in 2024.

One year later, the bank said it launched more than 60 AI solutions across multiple lines of business and planned to have AI chatbots integrated into seven different lines of business by the end of 2025.

The latest agentic AI model is the first step TD is taking to use the technology for its real estate secured lending, it said.

TD’s aim is to embark on an enterprise-wide AI strategy which has the mission to “deliver faster, simpler and more personalized banking experiences.” The bank has set a goal to generate $1 billion in annual value from AI in the coming years.

"We're building a hybrid future where our colleagues and AI work together to help our clients get to a 'yes' faster on some of their most important decisions,” Luke Gee, TD’s chief analytics and AI officer, said.

TD is emphasizing responsible use of AI, he added, to maintain the trust of its clients and employees.

For example, the bank plans to have the integration of agentic AI overseen by its Trustworthy AI team, which judges AI models on factors such as privacy, security, fairness and accountability. The team validates the AI models through a series of checks and balances before it comes into contact with humans, TD said, and continues to monitor the models once deployed.

TD’s Trustworthy AI team was named the Best Responsible AI Program or Initiative in North America by DataIQ in 2025.

TD plans to introduce agentic AI into every step in the real estate secured lending process, it said, and to explore ways to incorporate the technology in its other businesses.