AI Will Handle 95% of Underwriting. The Deals That Matter Live in the Other 5%.

By Shaun Ashkenazy, Founder & CEO, Lendyx

There is a lot of capital entering private lending right now, and the environment is shifting in ways the industry has not seen before. Geopolitical and economic pressures are moving valuations, construction costs, and investor behavior all at once. At the same time, every lender in the country is being told that AI is about to rewrite how loans get made.

Both things are true. But the technology conversation skips over the thing that actually determines whether a lender performs through a cycle: the discipline underneath it. So before I tell you where I think AI takes this industry, let me walk through how we lend. The two are connected.

Private Lending Was Never Meant to Replace Banks

What we do does not replace traditional financing. It provides what traditional financing was never built to provide.

There is an entire economy of real estate investors and developers whose businesses revolve around building homes, fixing them, selling them, or renting them. Banks typically do not finance construction and bridge loans. Some can be competitive on rate, but their risk tolerance and structure are not designed for fast, short-term lending. Almost none are built for what a developer actually needs mid-project.

That gap is where we operate: higher leverage, faster closings, and more flexible underwriting. We do not require tax returns or W-2s, because these are business-purpose loans made against real estate, not consumer mortgages.

The math on speed is not subtle. A bank may take 45 to 60 days to close a loan. We close in about 10. Draws matter just as much, because a developer waiting on funds is a project standing still, so we release draws in two to four days. In a market where rates are higher, buyers are constrained, and margins are tighter, speed and flexibility are not conveniences. They are what make deals pencil.

Underwriting Starts With the Borrower, Not the Spreadsheet

Construction costs, tariffs, and mortgage rates all move, and no one can predict where they land, especially on construction loans that can run 24 months. So we underwrite the thing that holds up when the market does not: fundamentals.

The borrower comes first. Experience, track record, and the ability to navigate a shift are what carry a project through two years of conditions nobody forecasted. We want to know they have done it before and can do it again.

Then we evaluate the market and the deal itself. Comps, appraisals, background checks, credit reports, submarket analysis, construction budgets, overall feasibility. All of it runs in parallel, which is how we get from file intake to closing in about 10 days without cutting corners.

We also decide fast, including the decision to pass. We see a high volume of deals and turn down plenty of them quickly. That volume is an asset. The more deals we evaluate, the sharper our judgment gets.

How We Protect Performance When the Market Moves

Our model is different from most of the market. We do not have traditional investors; we sell our loans. That structure means our protection has to come from deal selection upfront, not portfolio management after the fact.

In practice, that looks like an average borrower FICO above 750 and a book weighted toward institutional-level sponsors with real track records, liquidity, and net worth. It means concentrating on prime markets with strong fundamentals: Miami, Fort Lauderdale, Boca Raton, West Palm Beach, parts of California, and Charlotte. We lend nationwide, but we are deliberate about where the weight of the book sits.

None of that is complicated. It is discipline applied before the wire goes out instead of after.

Now, About AI

AI will have a major impact on financial operations, and underwriting is where it hits first. Underwriters are human, and AI can significantly reduce human error. We are investing heavily here, including a loan-sizing system that pulls comps, analyzes markets, and structures deals in a fraction of the time it takes today. That cuts the time from intake to term sheet and gets borrowers faster, more accurate answers.

But AI will not replace human decision-making, and here is the part most of the industry gets wrong: the real value in lending comes from the deals that fall outside standard guidelines.

Most deals, and often the best ones, require creativity. AI can analyze numbers. It cannot read the nuances of a deal: the sponsor's situation, the structure that makes a hard deal work, the judgment call on a project that looks wrong on paper and right in person.

I see AI handling about 95% of the work. The remaining 5% is where lending actually happens. Companies that use AI well will have a clear advantage, and companies that do not will fall behind. But the winners will be the ones who spend the efficiency they gain on more human judgment for that 5%, not less.

What Borrowers Are Actually Buying

Alignment of interest is natural in this business. If the borrower is successful, we are successful. If they are not, we are not. That accountability is why reputation and relationships are the two things we protect above everything else.

Day to day, that means execution. Answering the phone. Closing when we said we would. Honoring the terms we committed to. Doing the diligence upfront so there are no surprises at the table.

Borrowers choose lenders on pricing and trust. They want competitive terms, but what they are really buying is a partner who closes and services the loan the way they said they would. Certainty of execution is the new pricing. That is what drives repeat business, and repeat business is the whole model.

The market will keep shifting. The technology will keep improving. The lenders who win will be the ones who let AI do the 95% and never lose their edge on the 5%.

‍