TD is relying on artificial intelligence to strengthen its mortgage and retail banking and tells investors that both have a $40 billion opportunity to win customer mortgages held on rival lenders, as well as significant efficiency gains from automated approvals, approvals and customer service.
CEO Raymond Chun formed the strategy to deepen relationships with existing clients. “In Canada, we have earned the trust of a huge customer base and led the top position of retail banks,” he said.
He also highlighted efficiency as a cornerstone of the program, telling investors that AI-driven speed and consistency will play a key role in customer satisfaction and long-term profitability.
“We are approving mortgages for hours rather than days. We are providing millions of customers with credit cards with data-driven insights. We produce reports in minutes with hours or days, and we respond to customers in just a few seconds, greatly reducing calls and waiting times.”
The bank noted that the AI ruling that enables AI support has reduced turnover time, increasing one-day mortgage approvals by 20% in the mobile expert channel.
TD also highlights TD Mortgage Direct, a digital channel that transforms online interest into an instant callback for experts. The channel has funded $4.6 billion and converted at four times the old process.
Now that 93% of regular transactions have been processed digitally, TD wants branches to focus less on bills and deposits, but more on mortgages, renewals and investments. Another 500 branch employees will be redeployed for home lending or investment roles and referrals to wealth consultants have risen 18%.
TD has also restructured its proprietary mortgage channels, allowing mobile experts and branch bankers to conduct complex transactions more intimately. The bank confirmed that the cooperation has shown results, with branches having three times as much funding as mortgages and an overall productivity increase of more than 40%.
How AI reshapes TD
TD uses Investor Day to use AI as the entire bank operating system, rather than accompanying projects. Leaders have achieved early victories from automation and predictive models in loans, cards, services and insurance.
- mortgage: With the increase in AI support rulings, the coverage cycle time has shifted from days to hours.
- Judgment: Once hours of documents and qualification comments are spent are compressed to a few minutes.
- Card: TD is leaning towards tendency and risk models to expand credit card pre-approval for existing major customers – “starting with “yes” while holding credit quality.
- Service Operation: The internal GPT-style tool is touching the time of the crashing handle in the center and then entering the branch.
- Insurance: AI Claims Management is expected to remove $40 million in annual fees, while TD’s pilot agency workflow is designed to resolve direct auto claims in 15 minutes or less.
Chun said these examples show how AI can reshape the customer experience, not just internal workflows.
He also highlighted the scale of TD’s ambitions, emphasizing that the bank’s investment in AI is not about incremental improvements, but about providing measurable benefits across the enterprise. “We target AI at a $1 billion annual value, and we’re increasing revenue by saving costs, half is half,” he said, noting the bank’s internal expertise on the 6th floor of the Toronto-based AI Research Lab and data data from 2,500 people and AI Engineering Bench.
Replace costs to resolve investors’ concerns
In addition to new products and customer-facing tools, executives also stressed the importance of restoring bank efficiency.
CFO Kelvin Tran noted that TD’s efficiency ratio is currently about 58%, and plans to push it up as spending savings continue. The bank saved $2.5 billion in savings for $2.5 billion, including $450 million from moving more transactions to digital channels and reshaping its branches. Tran told investors that 2025 will mark a peak in expense growth, with efficiency expected to be standardized to the 1950s in the medium term.
Executives stress that AI is not only for efficiency, but also for tightening supervision. This becomes especially important when analysts question whether scaling AI can actually enhance operational or credit risk in so many processes.
Management responded to automation the contrary: by removing manual variability, reducing risks and enhancing fraud detection and adjudication consistency. As Chun said, the gains are “not risky, but risky by eliminating manual touch.”
Going forward, TD aims to have a return on equity of 13% in fiscal 2026 and 16% in 2029, with the main drivers of mortgage growth and fee revenue.
Visited once today, visited once today
Agent AI AI Big Bank Investors Day Kelvin 6th Layer Mortgage Loan Raymond Chun TD TD Bank TD Mortgage Direct Technology
Last modified: September 30, 2025
