How New Silver uses AI to speed up real estate lending
Alex Shvayetsky is Co-Founder and Head of Risk at New Silver, which provides technology-driven financing for residential real estate investors. This conversation explores how AI-assisted intake speeds up initial screening, why complex credit decisions remain human-led, and how borrowers are arriving better informed.
Start with one repetitive workflow, such as document intake or initial eligibility screening, and measure its cycle time and error rates. Keep human accountability for complex credit decisions.
What’s the quick origin story of your business, and what makes what you do genuinely different from the other options your customers are looking at?
New Silver was founded in the West Hartford, Connecticut area in 2018 and 2019 to close a gap I saw in real estate investment lending: investors with solid fix and flip or construction deals were losing properties because traditional lenders moved too slowly and buried them in rigid documentation. We built the company around a technology-driven process that can review a borrower quickly, often with a five-minute online approval, and fund in as little as five to ten days.
What genuinely separates us is that our speed comes from a standardized, data-driven risk playbook rather than from skipping checks. We encode borrower, property, rehab budget, after-repair value, and loan-to-value or loan-to-cost criteria into initial screening, then route complex or higher-risk files into tiered human review with documented exceptions, so investors get fast decisions without sacrificing credit discipline.
Where does most of your new business come from today, and where do you wish it came from?
Most of our new business today comes from real estate investor referrals, broker and strategic partner relationships, and repeat borrowers who have already closed with us. That channel works because experienced investors and brokers can vouch for our speed and reliability, and repeat clients come back when they have another project that needs fast capital.
I wish a larger share came from direct inbound investors who arrive already educated about DSCR, loan-to-value, rehab budgets, and timelines, and who can self-qualify before they contact us. That would let our team spend less time on basic education and more time structuring deals, and it would make our growth less dependent on partner relationships.
What’s one thing you’ve actually handed over to AI in the last year, and one thing you tried it on and went back to doing the human way?
We have handed over initial document intake and first-pass eligibility screening to AI-assisted tools, so borrower submissions can be read, categorized, and checked against basic criteria before a human reviewer opens the file. That removes repetitive sorting work and helps clean files reach the right reviewer faster.
We tried using AI to make final credit and exception judgments on complex files, and we went back to human review for that. The edge cases in construction draws, unusual property conditions, and borrower explanations require context and accountability that a model could not reliably provide, so we kept final risk decisions with trained underwriters.
Have you noticed a change in how customers find you or what they already know before they reach out? Is anyone arriving through AI assistants like ChatGPT yet?
Yes, customers arrive better informed than they did a few years ago. More investors already understand DSCR, loan-to-value, after-repair value, and funding timelines before they contact us, and they ask sharper questions about draw schedules and eligibility instead of starting with basic definitions. That shift makes the first conversation more substantive.
We are seeing some early arrivals through AI assistants like ChatGPT, but it is still a small and hard-to-measure share compared with referrals and search. Those prospects often come in with a shortlist of lenders and specific criteria, so we treat AI-assisted discovery as a real but emerging channel rather than a primary source today.
If you were advising someone in your industry on all of this for the next 12 months, what would you tell them to actually do, and what would you tell them to ignore?
I would tell them to write down their underwriting criteria, decision rights, and exception process, then clean up their loan data so any AI tool has accurate inputs to work from. The practical first step is to automate one narrow, repetitive workflow such as document intake or initial eligibility screening, measure cycle time and error rates, and keep human accountability on final credit judgments.
Thank you to Alex Shvayetsky and the team at New Silver for sharing what actually worked with Leaders Perception readers.
Want to share your playbook?
Leaders Perception publishes founder and operator interviews every week. A few questions, about five minutes, no calls. It is free and everyone we accept gets published.