TD introduces agentic AI to streamline mortgage applications
Payments Canada prepares instant payment rollout with phased access for banks, FinTechs, and payment providers.
TD Bank Group introduced its first agentic AI model, designed to streamline and automate the pre-adjudication process for both mortgage and Home Equity Line of Credit (HELOC) applications within its Real Estate Secured Lending operations. Developed by TD’s Layer 6 AI research centre in partnership with internal risk and technology teams, the model generates concise application summaries for underwriters in minutes rather than hours, reducing manual tasks and enhancing efficiency across the bank’s end-to-end lending workflows.
Key highlights
Agentic AI automates pre-adjudication: The model employs autonomous AI agents with generative language capabilities to classify documents, extract key data, calculate borrower income and validate it against policy rules before producing a concise memo for underwriters. Early tests show it delivers more accurate results while dramatically reducing adjudication time and unit costs.
Part of a wider AI strategy: TD intends to scale the technology across all steps of its real‑estate secured lending (RESL) journey, from document submission to funding. The bank aims to combine AI with human expertise to simplify the mortgage and home‑equity line of credit experience and generate around $1 billion in annual value from AI initiatives in coming years.
Oversight and Control: TD’s Trustworthy AI team assesses the model for privacy, security, fairness and accountability, applying strict checks before deployment and monitoring its performance afterwards. This governance framework is designed to uphold client trust as autonomous agents take on more complex tasks.
Roadmap Beyond Mortgages: The launch marks the first step toward transforming TD’s RESL operations. The bank has mapped every stage of the lending process and is exploring uses for agentic AI in other divisions, signalling that autonomous AI agents may soon influence multiple areas of TD’s business.
What this tells us about the ecosystem
The deployment of agentic AI by a major Canadian bank underscores the growing adoption of generative AI in financial services. Institutions are moving beyond AI pilots to embed autonomous agents in core lending workflows, aiming to boost speed, accuracy and customer experience. This shift highlights a broader trend toward hybrid models where AI handles routine analysis while human experts manage complex judgement calls, with strong governance frameworks needed to manage risk and maintain public trust.
