Sun West’s AngelAi Slashes Mortgage Decline Rates and Expands Fair Lending Access

Research into mortgage applications processed through AngelAi at Sun West Mortgage Company reveals that AI-driven processing significantly reduces decline rates across all ethnic demographics compared to national benchmarks. For fintech professionals, this suggests that high rejection rates may stem from the operational costs of complex applications rather than inherent credit risk, potentially unlocking $230 billion in annual loan volume.

What was announced

The findings, published in the white paper "Operationalizing Fair and Accessible Lending with AI," compared 2025 AngelAi outcomes against national Home Mortgage Disclosure Act (HMDA) data. The study found that decline rates were 24 to 29 percentage points lower than national averages across every ethnicity analyzed. Specifically, Asian applicants saw a 4.01% decline rate with AngelAi versus 32.98% nationally, while Black applicants recorded 18.55% compared to 44.32% nationally. Hispanic and White applicants also saw significantly lower rejection rates at 13.73% and 9.24% respectively.

Crucially, the research indicated that these lower decline rates did not lead to a material increase in delinquency, default, or foreclosure. The platform utilizes a patented Transactional Language Model (TLM) which, unlike standard Large Language Models (LLMs), follows defined rules to ensure consistent and auditable results. This automation handles document intake, income analysis, and condition management, reducing the manual labor typically required for borrowers with multiple jobs or irregular income. The report follows a $100 million investment in Celligence, the developer of AngelAi, from real estate finance firm Mortgage Treasury and Japan-based Matsunaga Construction. To date, AngelAi has registered over 415,000 users and facilitated more than $40 billion in transactions, with plans to expand the TLM technology into other financial services and international markets.

The companies involved

Sun West Mortgage Company is a major player in the mortgage sector, led by CEO Pavan Agarwal. The company has focused on integrating advanced technology into the lending lifecycle to address systemic inefficiencies in how loans are processed and approved. The technology itself, AngelAi, was developed by Celligence, a financial technology firm that specializes in rule-based AI systems designed for high-stakes financial transactions where probability-based generative AI is insufficient.

The project has attracted significant international backing, notably from Mortgage Treasury, a real estate finance firm, and Matsunaga Construction. Based in Japan’s Greater Tokyo Area, Matsunaga Construction is a construction and infrastructure company that has recently identified itself as a lead investor in the $100 million funding round for Celligence. The involvement of a Japanese construction firm highlights a strategic interest in diversifying assets away from yen-dominated holdings while backing technology that addresses demographic shifts and labor shortages. Celligence intends to use this capital to scale AngelAi’s user base and explore international markets, moving beyond the US mortgage sector into broader financial applications.

What this means

This data challenges the long-held assumption that high decline rates among certain demographics are purely a function of credit quality. By reducing the "manufacturing cost" of a mortgage from $5,000 to $125, Sun West has effectively commoditized the processing of complex files—such as those from borrowers with multiple jobs or irregular income. This puts immense pressure on traditional lenders whose high overheads force them to prioritize "clean" applications to maintain margins. If AI can prove that complexity does not equal risk, the industry faces a reckoning: the "credit gap" may actually be an "efficiency gap" that legacy systems are no longer equipped to bridge.