A new mortgage technology startup is betting that borrowers will increasingly turn to artificial intelligence to find a loan originator, and it wants to make sure AI knows whom to recommend.
LenDb.ai, launched Thursday by venture studio Draper Goren Blockchain, has assembled a free database of more than 224,000 U.S. loan officer records designed for use by AI assistants, including ChatGPT and Claude.
The company is also taking an unusual approach to its own operations. An AI agent named Andrew handles much of the day-to-day work, from maintaining the database to writing software and communicating with lenders, under the supervision of human executives.
The database is intended to help borrowers find originators based on location, licensing, loan specialties, and languages spoken. Someone looking for a Spanish-speaking loan officer in Phoenix who handles VA loans, for example, could use a connected AI assistant to search for a match.
The concept introduces another potential source of borrower inquiries at a time when visibility online often comes at a cost.
"At least ten of our companies work in the proptech space, and over and over we ran into the same problem: finding a loan officer is an extremely biased, pay-to-play game," said Alon Goren, founding partner of Draper Goren Blockchain and board co-chair of LenDb.ai. "The lists borrowers see are ranked by who paid, not by who's right for them."
A Different Approach To Finding Borrowers
Unlike paid lead-generation platforms, LenDb says lenders cannot purchase placement or pay to improve their rankings. The company also says it does not sell leads or collect information about which LOs borrowers contact.
Its database is available through a free application programming interface (API) and Model Context Protocol (MCP) server, allowing compatible AI assistants to retrieve loan officer information.
The company's Nov. 3 database snapshot lists 224,530 loan officer records, including 149,751 identified as active. Of those, 48,228 have full profiles drawn from lender websites. The remaining records are largely licensing-based entries, which the company says are excluded from search indexes until additional information is available.
A licensing record establishes a person's presence in the database, while a more complete profile can provide the specialties, languages, and professional background needed to match a specific borrower request.
Loan officers can claim and correct their profiles, and lenders can submit information about their teams. LenDb says those contributions are identified as self-reported.
Lit Financial is among the early participating lenders.
"Our loan officers are the reason borrowers choose Lit," said Tim Barry, founder and CEO of Lit Financial. "LenDb makes it easy for their experience, specialties and own words to show up when someone asks an AI for help, and it was simple to get our team on it."
LenDb's published disclosures identify an ownership interest in Lit Financial held by LenDb's owner. The company says that relationship does not affect rankings or provide Lit Financial with preferential referrals.
An AI Agent Running Daily Operations
LenDb's management structure is another departure from the typical mortgage technology startup.
Draper Goren assigned daily operations to Andrew, an AI agent developed from the firm's internal assistant, Brobot.
Andrew maintains the data pipeline, processes licensing information, handles corrections, writes software, and communicates with lenders and partners. The agent also reports to a human board co-chaired by Goren and Draper Goren co-founder David Bleznak.
Although the company refers to Andrew as its AI CEO, its published disclosures identify Goren as the human CEO of LenDb.ai Inc. Andrew holds no corporate office and cannot sign contracts or make legally binding commitments.
"Since LenDb is built for bots, we decided a bot should run it," Goren said. "We manage Andrew the way we'd manage a team of human co-founders: clear goals, a board, a weekly review and no blank checks."
The arrangement gives the venture studio a way to test how much of a startup's routine work can be delegated to AI while keeping financial and legal authority with people.
What It Means
LenDb offers another potential way to reach borrowers without paying for leads. But getting found through AI may depend on how accurately an originator's specialties, licensing, and experience are represented online.
LenDb draws that information from public licensing records and lender websites. The company acknowledges that some details may be outdated or incorrect, and its rankings do not currently account for regulatory or disciplinary histories. Borrowers are directed to NMLS Consumer Access to verify licensing.
While much of the mortgage industry's AI investment has focused on underwriting, processing, and operational efficiency, LenDb is targeting the front end of the business: how borrowers find an originator.
