The AI-powered underwriting system developed by Better.com can evaluate mortgages in minutes, even seconds. This is incredible speed that can significantly compress one of the most time-consuming parts of the lending process.
On March 5, Better announced the launch of a conversational credit decision engine for mortgage and home equity loans built in partnership with OpenAI and available on ChatGPT. The new application allows lending teams to connect directly to Better’s Tinman underwriting engine from within ChatGPT to generate loan evaluations, allowing lenders to approve loans faster and streamline underwriting workflows.
With this release, Better is expanding access to the Tinman AI platform across the financial services industry through a custom Model Context Protocol (MCP) connector developed by the company’s engineering and AI teams in collaboration with OpenAI.
humans in the loop
According to the company, the Tinman AI platform currently manages more than 1,200 different workflows across the mortgage process, including underwriting. In many cases, the system can complete the entire underwriting process end-to-end. Licensed insurers step in when AI detects fraud or scenarios that require professional judgment.
Vishal Garg, CEO and founder of Better.com, told Inman in an email that “a licensed professional will be brought in to review any discrepancies detected and perform a final review of the file, which will require a licensed underwriter prior to financing.”
Vishal Garg | Photo Credit: Better.com
These human reviews serve another purpose as well. Feedback is used to continuously improve the platform’s machine learning models.
47 seconds underwriting request
Better said the system can underwrite some loans in as little as 47 seconds, a number that reflects highly standardized loan scenarios. “While some loans have been underwritten in as little as 47 seconds, the typical median loan underwriting timeline on the Tinman Credit Decision Engine app is 2 minutes and 24 seconds,” Garg said.
This compares to the traditional mortgage underwriting process, which often takes 15 to 21 days. The company says traditional loans made to eligible W-2 borrowers tend to be processed fastest. More complex loan files may take several minutes and may require a final review by a licensed underwriter.
“Some loans can take up to five minutes and others require handing over to a licensed underwriter for final review, adjudication and approval,” Garg said. “AI makes this easier by doing all the manual work.”
Enterprise clients of the Tinman AI platform, such as mortgage companies, fintechs, and loan officer teams, will have access to the Tinman Credit Decision Engine app if they have access to ChatGPT Enterprise.
Tools to “supercharge” your loan officer
If underwriting schedules continue to shrink, the role of loan officers could gradually shift from administrative-heavy processing tasks to providing advice to borrowers. Better argues that automation allows loan officers to focus on helping buyers make the biggest financial decisions of their lives.
“Buying, owning and maintaining a home is a fundamentally human process,” Garg says. “This tool allows our team of loan officers to engage in the process as needed and eliminates the complexity of paperwork that would be better left to a machine.”
Better’s loan officers, processors, and operations teams are already using this integration to reduce underwriting escalations, prevent rework by validating key process steps before taking action, and accelerate onboarding by giving new hires immediate access to Tinman’s organizational knowledge.
“This is a tool that will greatly enhance loan officers, allowing them to spend more time on higher-value work, like helping people find homes,” Garg said.
The company also points to productivity gains from previous versions of its AI platform. In a case study on lender NEO, Better said the Tinman system improved loan volumes per loan officer by more than 90 percent within six months of implementation.
“This is huge in the context of consumer mortgage origination, which is part of the mortgage industry, where the average loan origination cost per loan is $11,700 and there has been almost zero growth in loan productivity per loan officer over the past 10 years,” Garg told Inman. “We look forward to further improvements with the addition of the Tinman AI app to ChatGPT.”
“30x better than human underwriters”
Making underwriting decisions quickly inevitably raises questions about accuracy and regulatory compliance. Better said Tinman’s platform is trained on nearly a decade of institutional mortgage intelligence and powers a machine learning-driven rules engine.
The Tinman platform also includes a “context graph” that maps individual mortgage processing and underwriting roles, mapping tasks, actions, rules, and decisions across more than $110 billion in reserve loans, more than 12 million customer call records, and billions of pages of credit, income, housing, and asset-level documentation.
The system achieves a critical defect rate of 6 basis points, which the company says is significantly higher than the industry average of 1.8%.
“This means the platform is 30 times better than human processors and underwriters at accurately underwriting loans according to investor guidelines for over 40 investors across a wide range of products that meet all the compliance and regulatory conditions that every mortgage loan must adhere to,” Garg said.
The “scary part” for mortgage loan professionals
Better’s Tinman AI platform represents a growing shift toward automation in mortgage underwriting, and the company’s claims are supported by compelling data. But some industry observers remain skeptical.
Colin Robertson, who runs the mortgage news website The Truth About Mortgage, said the headline-grabbing claim of 47-second approvals may be more about marketing than a fundamental change to lending schedules.
“Given the company’s emphasis on 47-second loan approvals, you can tell this is mostly a publicity stunt,” Robertson told Inman. “Who cares if it takes less than a minute to figure out you are?” [conditionally] The whole process still takes 30-45 days and it gets approved?”
Still, Robertson said tools like Better’s conversational underwriting engine are a step toward a more automated mortgage pipeline. “On the other side of the coin, the more companies that integrate AI, the closer we get to an automated and streamlined mortgage experience from start to finish,” he said.
Robertson noted that this shift could have a significant impact on the mortgage workforce if automation continues to expand into underwriting, loan processing and document verification.
“The scary thing for anyone working in this industry is that the goal is clearly to replace humans with automated workflows, no different than Waymo getting rid of drivers,” Robertson added.
Email Nick Pipitone
