AI reshapes secondary market pricing

National Mortgage News reported on May 22 that mortgage executives are preparing for artificial intelligence to play a larger role in how mortgage loans are priced and traded after origination. The report said lenders and investors see AI as a way to analyze large pools of loan-level data, identify pricing drivers that humans may miss and reduce execution surprises in the secondary market. The discussion surfaced as the Mortgage Bankers Association held its Secondary and Capital Markets Conference in New York from May 17 to May 20. MBA said the event gathered market makers, agency leaders and mortgage executives to discuss current capital-markets conditions and technology. ### Where would AI show up first in mortgage trading? National Mortgage News said the first use case is not a fully automated trading desk but deeper analysis of loan files before trades are executed. The publication said executives described AI as a tool that can scan large data sets and isolate granular characteristics inside individual files that affect how investors value loans. That work now depends heavily on human review, spreadsheets and trader judgment. (nationalmortgagenews.com) The same report said firms expect that added file-level visibility to improve pricing precision and help traders understand why one loan pool may clear differently from another. In secondary markets, small differences in borrower profile, documentation or property characteristics can alter bids, hedge performance and pull-through assumptions, according to the report. ### Why are lenders focused on “granular” characteristics? (nationalmortgagenews.com) Loan-level pricing already depends on details such as credit profile, loan size, occupancy and product type, but National Mortgage News said executives expect AI to widen the number of variables that can be evaluated at once. The report said that could help firms uncover patterns across thousands of files rather than rely on broader averages for whole pools. (nationalmortgagenews.com) MBA has separately highlighted data quality and connected-data workflows as a live issue for capital-markets and mortgage operations in 2026. Its conference schedule included sessions on liquidity pressures, regulatory uncertainty and the need for clean, connected data, underscoring that firms see better data infrastructure as a prerequisite for more automated decision-making. (nationalmortgagenews.com) ### Does this mean AI will actually place trades? National Mortgage News said some mortgage leaders went further and said AI agents could eventually execute trades themselves. The report framed that as a future possibility tied to AI’s ability to uncover and analyze loan-level information at scale, rather than as a system already in broad production across the market. The same article did not describe a marketwide rollout timetable, and MBA’s public conference materials do not indicate that autonomous trading is now standard practice in mortgage secondary markets. (mba.org) Instead, the available material points to a nearer-term phase in which firms use AI to support pricing, execution and data review while humans remain responsible for trading decisions and risk controls. That is an inference from the conference agenda and the National Mortgage News report. (nationalmortgagenews.com) ### Why would brokers care about a back-office pricing story? Brokers rarely interact directly with the secondary desk, but National Mortgage News said the practical effect could show up in lender pricing consistency and lock execution. If AI helps lenders identify pricing drivers earlier and reduce surprises when loans are sold, brokers could see fewer unexplained price changes and faster responses on locks, extensions or exceptions, according to the report. (nationalmortgagenews.com) That could affect how wholesale lenders present themselves to broker partners. In that pitch, the differentiator would not be AI by itself but whether technology produces steadier pricing and cleaner execution for brokers trying to manage borrower expectations in a volatile rate market, the report said. ### What happens next for this technology push? MBA’s next publicly listed industry milestones include its June 1-4 MISMO Summit, where standards and technology integration are on the agenda, and later 2026 mortgage conferences focused on compliance and operations. (nationalmortgagenews.com) National Mortgage News has also continued broader coverage of AI in mortgage workflows this year, including underwriting and lender use cases, suggesting secondary-market pricing is part of a wider industry push to apply AI to data-heavy parts of the business. (mba.org)