TORONTO–While AI has been deployed across a host of back office functions, on the consumer-facing side its promise is increasingly being seen in mortgage lending, where lenders are promising mortgage approval decisions in as little as 47 seconds, reporting that up to a third of inquiries are now being handled by chatbots, and slashing underwriting time to just minutes.
Toronto-based TD Bank Group said it has also deployed its first agentic artificial intelligence system in mortgage lending, reducing the time required to prepare applications for underwriting from an average of roughly 15 hours to less than three minutes.
According to a statement from TD Bank, the new AI model automates mortgage pre-adjudication — the process that occurs before a human underwriter reviews an application.
The bank said the system classifies borrower documents, extracts and validates financial information, calculates income, performs policy and consent checks, identifies discrepancies and generates a summary memo for underwriters.
Developed Internally
TD said the model was developed by Layer 6 AI, the bank’s internal artificial intelligence research and development center, in collaboration with technology, data, lending and risk management teams.
The bank said the initiative is intended to accelerate decision-making during the homebuying process while maintaining human oversight.
Mohit Veoli, TD’s senior vice president of real estate secured lending, said in the statement that the technology is designed to provide “confident decisions earlier in the homebuying process.”
‘Hybrid Future’
Luke Gee, TD’s chief analytics and AI officer, said the bank is pursuing what he described as “a hybrid future where our colleagues and AI work together to help our clients get to a ‘yes’ faster.”
TD, which said it serves 28.1 million customers and ranks as North America’s sixth-largest bank by assets, described the mortgage launch as the first implementation of a broader agentic AI strategy.
According to the bank, it has mapped the entire mortgage and home equity line of credit process — from document collection through loan funding — and plans to introduce agentic AI capabilities across additional stages of the lending lifecycle.
TD also said it is evaluating similar applications across other business lines.
‘Hyper-Fast Closings’
Meanwhile, Palantir Technologies, the AI company that is backed by investor Peter Thiel, among others, has become the latest to enter into a partnership aimed at AI-backed mortgage lending decisions.
RocketClose is also touting hyper-fast closings.
Palantir said it is partnering with Moder to develop an artificial intelligence-powered mortgage operations platform, with Freedom Mortgage serving as the initial pilot customer.
The companies said the platform will combine Palantir’s data and AI capabilities with Moder’s mortgage industry expertise to streamline lending operations, with the stated goals of improving affordability, reducing borrowing costs and expanding access to homeownership.
The system is built on Palantir’s Ontology technology, which the companies said enables an “agentic AI” framework capable of integrating with existing systems of record. The platform translates lending guidelines and operational policies into configurable, testable and auditable rules, allowing mortgage teams to execute processes with greater speed, accuracy and scalability, the companies said.
The Early Developments
Early deployments at Freedom Mortgage have already supported several operational functions, the companies said, improving processing times and accuracy for both staff and borrowers, the companies reported.
Michael Middleman, chairman of Moder, said in a statement the collaboration has the potential to reshape mortgage operations, while CEO Erik Anderson said the combined capabilities are already producing measurable improvements in efficiency and the customer experience.
47-Second Mortgage
As the CU Daily reported here earlier, numerous AI-backed initiatives in mortgage lending have been announced, with the mortgage platform Better saying it has partnered with OpenAI to launch an app within ChatGPT that the companies said will dramatically reduce the time it takes to underwrite a mortgage or home equity loan to as little as 47 seconds.
The app takes Better’s mortgage engine and combines it with OpenAI’s models to speed up the underwriting process for loan officers working at banks, mortgage brokers and fintech firms, Better CEO Vishal Garg told CNBC in an interview.
In addition, earlier, in 2025, Fannie Mae partnered with Palantir on fraud detection initiatives, while Better previously used Palantir technology to launch its Tinman Marketplace in 2022.
Automating Interaction
Meanwhile, Better.com reported it is using generative artificial intelligence to automate a growing share of mortgage borrower interactions, with the company’s AI-powered loan agent now handling nearly 100,000 mortgage-related calls per month, according to ElevenLabs.
The AI agent, known as Betsy, resolved 35.5% of borrower inquiries in 2025 without human intervention, according to a case study released by ElevenLabs, whose voice technology powers the platform. The report said Betsy placed approximately 1.89 million inbound and outbound calls during the year.
The technology reflects a broader shift within financial services toward voice-based AI systems designed to streamline consumer transactions, according to analysis from PYMNTS.
Automating Conversations
Better.com said the system allows the mortgage lender to automate routine borrower conversations while maintaining around-the-clock customer support. HousingWire previously reported that Better CEO Vishal Garg said Betsy can perform rate calculations across more than 26,000 mortgage product and investor configurations.
According to ElevenLabs, Better initially relied on a speech-to-speech AI model before moving to a modular system combining speech-to-text, a large language model and text-to-speech technology through ElevenLabs Agents. The company said the new architecture reduced latency and improved compliance controls.
The case study said the AI system generated a 41% reduction in average loan origination costs in 2025 and doubled Better’s lead-to-lock conversion rate.
Margin Pressures
The cost reductions come as mortgage lenders continue to face margin pressure. According to Freddie Mac, loan origination costs in recent years climbed 35%, while the average retail-only mortgage lender lost roughly $600 per loan.
The companies said the new platform could help mortgage providers reduce operational costs, accelerate loan processing and increase lending capacity through more automated and precise workflows.
Rocket Close Rockets Approval Time
Separately, Detroit-based Rocket Close said it has cut mortgage document processing times from hours to minutes after deploying a generative artificial intelligence system developed with Amazon Web Services.
The company, which processes roughly 2,000 abstract document packages each day, previously relied on manual workflows that could take up to 10 hours per package as volumes increased, HousingWire reported. Each package averages about 75 pages and includes detailed legal and financial records tied to property ownership and lending.
With the new AI-driven system, processing time has been reduced to less than two minutes per package while maintaining about 90% accuracy in document classification and data extraction, according to HousingWire.
Humans to Remain Involved
Nathan Schrauben, Rocket Close’s chief information officer, said human oversight remains a key part of the process.
“The human will step into almost all of the transactions, whether it’s verifying the data or looking at the exceptions,” Schrauben said, according to HousingWire. “That ensures we’re doing the right thing for our clients and making sure they have the proper homeownership rights to the property.”
Betsy conducts eligibility checks, discusses pricing and assists borrowers with rate locks, while human loan officers focus on more complex borrower situations and underwriting decisions, according to the report.
ElevenLabs Head of Agents Growth Lauren Rothwell said deploying AI voice systems in regulated industries such as financial services requires accuracy and compliance in addition to conversational capability.
Earlier this year, ElevenLabs raised $500 million in funding at an $11 billion valuation, reflecting growing investor interest in AI voice infrastructure for enterprise applications.
