Why We Built a Client Portal for HOA Loans

The AI-enabled client portal we launched in June 2025 started with a phone call I will not forget. A board treasurer in Florida called me from her car. She had just taken a half day of paid time off, which she could not really afford, to drive to her management company's office and fax three governance documents to a lender's underwriting team. It was 2023. She was faxing. From a real fax machine. Because the lender's portal would not accept the file format her management company exported in. I sat in my office and thought, this cannot still be how we do this.

That call is the reason we built the portal. I have spent 30+ years in commercial banking and 20+ years serving on HOA boards, and I have watched the same scene play out in different cities and different decades. A volunteer treasurer, who has a day job and a family and zero training in commercial real estate finance, gets asked to assemble a 40-page loan application packet. They struggle. The deal slows. Sometimes the deal dies because the treasurer ran out of time before the lender ran out of patience.

The Portal We Wished Existed When We Started

If you had asked me when we started the firm in 2016 what HOA lending needed, I would have said the same thing I would say now. The technology should serve the borrower, not the lender. The borrower in our world is a volunteer doing this on top of a real job. Everything else flows from that.

The technology went the other way. Lenders built portals optimized for their own underwriters. They asked treasurers to fill in forms designed for someone who knew the difference between operating reserves and replacement reserves on the first try. Boards uploaded what they had, in the format they had it, and got rejected emails asking for the same documents in slightly different forms.

We sat in those meetings. We watched boards quit halfway through. We saw projects get downgraded to special assessments not because the loan math did not work, but because the loan process was too painful for a volunteer to finish. That outcome is a failure of the industry, not the treasurer. The treasurer is doing exactly what we should expect a treasurer to do: trying their best in spare hours.

The Design Principle

One sentence governed every decision. The portal works for the volunteer treasurer, not the lender's underwriter.

That sounds obvious. It is not. Many other portals in this space started with the question, what does the lender need? Our portal started with, what does the treasurer have? Treasurers have a reserve study in PDF, financials exported from their accounting software, governance documents sitting in a shared drive, and 30 minutes between dinner and putting the kids to bed. If the portal cannot work in that environment, it is the wrong portal.

So we built it to accept what people actually have. PDFs in whatever shape. Spreadsheets in whatever format. We let the AI do the work of extracting and standardizing on the back end. The treasurer should not be cleaning data for a lender's underwriter. We should be cleaning the data, on behalf of the treasurer, before the lender ever sees it.

What We Deliberately Did Not Automate

This is the part that matters more than the rest. We did not automate the lender match.

Many AI vendors in adjacent markets are pitching full automation, end-to-end, no human in the loop. We considered it. We rejected it. The reason is simple. The judgment call on which lender is right for a specific community is not a pattern-matching problem. It is a relationship and history problem. We know which lenders are more likely to look past a temporary delinquency spike when there is a clear cause behind it, such as a single owner's estate. We know which lender's portfolio is full and which one is hungry. We know which lender treats Florida coastal properties differently than mountain west garden-style communities. None of that lives in a structured database. It lives in phone calls.

Ben and I, along with our team, make that call by hand on every deal. The AI prepares the package. Humans choose the lender. We don't underwrite loans, make credit decisions, or lend money. That division of labor is intentional, and we will not change it, because the moment we automate that decision is the moment the value of an advocate for the association disappears.

The Honest Argument

Here is the line I am willing to defend in any room. AI in HOA lending is only as good as the people behind it.

A client portal in the hands of a single lender just makes it easier for that lender to underwrite their own pipeline. The borrower gets a faster yes or no, but they get only one lender's opinion. A client portal in the hands of a captive broker (paid by a single lender, dressed up as independent) just routes you to the same place faster. A client portal in the hands of an advocate for the association paid only on closing changes the math. Now the technology is working for you, because that advocate can only get paid if the deal closes.

That economic structure, no upfront cost and payment at closing, is what makes the portal worth using. Without it, you are just feeding documents into a tool that serves someone else's underwriter. With it, the AI is doing what AI is genuinely good at, which is reducing the grunt work that wears volunteer treasurers down before the real conversation ever starts.

What This Means for the Treasurer Reading This

If you are a board treasurer reading this on your lunch break, what I want you to know is that the portal exists because somebody on our team picked up the phone when a Florida treasurer was crying in her car about a fax machine. We built it for her. We built it for you.

If your board is starting to think about a loan, the first step is a conversation. Portal access comes once we are working together on a live file, so start by getting in touch and we will tell you what we see and what the lender network is likely to say. The conversation costs nothing, because we get paid only if the loan closes.