Key facts
- The average cost to originate a mortgage loan remains stubbornly around $11,000.
- Compensation accounts for roughly two-thirds of the direct cost to originate a loan.
- Technology spending accounts for about 4% of the direct cost to originate a loan.
- AI-native operating models could reduce the cost to originate a loan to $8,090.
- AI adoption among lenders is projected to rise from 15% in 2023 to 60% in 2025.
The average cost to originate a mortgage loan has remained high, around $11,000, despite significant technology investments and the availability of automation tools. While some technology providers claim substantial cost reductions per loan, these benefits have not translated into a clear, sustained decrease in overall industry costs. Compensation consistently accounts for about two-thirds of the direct cost to originate, with technology making up only about 4%.
Industry experts suggest that a lack of process discipline is a primary reason technology has not meaningfully reduced origination costs. Employees often ignore or work around available technology if a specific business process is not enforced. However, the adoption of AI-native operating models, which can read files, balance disclosures, and route exceptions to staff, shows promise. One top 10 lender's deployment of SnapDocs reportedly reduced quality control review time by 71% and increased closing team throughput by 65%.
Modeling suggests that AI-native models could reduce the cost per closed loan to approximately $8,090, with a 56% reduction in expenses below the loan officer. Technology expenses are expected to rise, but the return on investment is estimated to be seven to ten times the spend. The remaining gap to top-quartile costs near $6,900 would require restructuring sales compensation, a strategic decision beyond technology investment. AI adoption is expected to grow significantly, from 15% in 2023 to 60% by 2025, but the aggregate industry impact has yet to fully materialize.