As murky sentiment looms over the housing market, Michael White sees a bright spot for tech to tamp down mortgage troubles
As the Federal Reserve weighs its next rate move — and Treasury Secretary Scott Bessent warns of a stagnating housing market — mortgages are once again in the national spotlight.
Against this backdrop, Multiply Mortgage, an AI-native mortgage platform and provider of mortgage employee benefit programs, is introducing two tools aimed at improving efficiency and easing costs for both consumers and lenders–Rate IQ which analyzes offers to reduce complexity for comparing mortgages, and Loan Sentinel which is aimed at reducing inefficiencies behind the scenes with documentation and coordination.
Future Nexus spoke with Multiply Mortgage co-founder and CEO Michael White about implementing AI into the mortgage process, the company’s mission, and expanding role of mortgage benefits in the workplace.
Can you explain the difference between being an “AI-native mortgage platform” and one that uses AI as a feature, possibly to enhance an existing platform? What makes you different?
Multiply Mortgage is built to be AI-native, not AI-assisted. We’ve built our origination platform from the ground up to best leverage AI, rather than augmenting the legacy workflows of a traditional lender.
This approach enables us to accelerate the mortgage process and cut costs, thereby removing inefficiencies faced by traditional lenders. Our vertical AI stack delivers operational efficiencies that translate directly into savings for borrowers—up to 1% off retail interest rates and an average of $5,700 in annual savings per employee.
Multiply Mortgage is operating in a highly regulated industry. What were the challenges you experienced as you sought to apply AI to this sector while appeasing any regulatory concerns?
We’re actually finding that the mortgage industry is quite accepting of AI technologies. According to Fannie Mae’s October 2023 Mortgage Lender Sentiment Survey, the AI mortgage lending market is expected to reach $10.4 billion by 2027, but it wasn’t that way from the start.
Early on, there was skepticism around transparency and explainability, especially given how heavily regulated the industry is. We addressed this by designing our AI systems to empower licensed professionals, not replace them. Every agentic workflow operates within existing compliance frameworks and keeps humans in the loop for decisioning oversight and auditing.
Our approach focuses on automating fulfillment, processing, and communication: areas that traditionally account for most of the industry’s cost and friction, yet fall outside of regulated credit determinations. This allows us to achieve substantial operational gains while maintaining full compliance with state and federal lending laws.
We also mitigate any regulatory risk that an employer might assume because we keep them completely out of the mortgage process. Once access is provided, they’re not a party to the loan and receive no transaction data, eliminating employer risk while protecting employee privacy.
In the past, some technologies that incorporated data extraction struggled with the lack of document uniformity and were challenged to identify and retrieve important information. How did you overcome this challenge?
Historically, document extraction relied on template-based approaches and traditional OCR tied to specific formats. Whenever documents changed, entire Data Science organizations would be dedicated to rewriting templates, a maintenance burden that was impossible to scale because in this industry, every institution formats documents differently.
LLMs solved this problem because they can classify documents and extract details based on content and semantics rather than rigid formatting rules. We utilize specialized models for various document types, cross-reference extracted data for validation, and surface low-confidence results to processors for review, rather than passing through potentially incorrect data.
The key difference is that LLMs understand context and meaning, not just position and format. We spend our time improving accuracy rather than maintaining brittle templates.
Without revealing any secrets, can you share a bit on how Rate IQ analyzes the various terms (which can be opaque) to deliver an apples-to-apples comparison?
Many loan comparisons focus on the rate, which does not capture the full cost and term considerations in a financial product as complex as a mortgage. Points, closing costs, and other terms can have a significant impact on the customer, so it is essential to standardize these terms to compare different loan options effectively.
APR is the standardized basis for comparison, but it doesn’t capture every consideration. To enable an apples-to-apples comparison with the market, we created Rate IQ to extract details, including rate, points, product type, and more, to price out loan options that are as close to the original as possible.
Share more about the philosophy behind your B2B2C model, one which is gaining popularity in fintech circles. Why is this model becoming more popular? Discuss the attraction to employers and employees, and why it makes sense for you to pursue it as a business opportunity.
Our B2B2C model is rooted in the belief that homeownership and employment are deeply connected: your job and salary determine your buying power, yet housing has been largely excluded from the employee benefits ecosystem. Employer distribution allows us to reach borrowers at the exact moment they’re ready to buy, while offering them exclusive pricing and concierge-level guidance through a trusted channel.
From a business standpoint, this model is both scalable and capital-efficient. Partnering with employers significantly reduces customer acquisition cost compared to direct-to-consumer channels and provides built-in credibility and reach — especially as HR and benefits leaders increasingly prioritize financial wellness. For employees, it’s an opportunity to access lower mortgage rates and better support; for employers, it’s a differentiated, zero-cost benefit that improves retention and satisfaction.
More broadly, the growing popularity of B2B2C models in fintech reflects a shift toward embedded distribution — meeting customers where they already make financial decisions, through trusted platforms and institutions rather than standalone ads or funnels. It’s a win-win model that aligns incentives across all stakeholders while efficiently building long-term customer relationships.
