A large customer requested automated underwriting capabilities within the Pillar platform, including pro formas, rent comps feeding into NOI projections, and the whole stack, to streamline deal analysis.
I told our biggest customer no. They threatened to cancel. I still said no. They wanted automated underwriting inside Pillar. Pro formas, rent comps feeding into NOI projections, the whole stack. On paper it made perfect sense. They were paying us five figures a year and it was the one thing they kept asking for. "Just add underwriting" sounds like a feature request. It's actually a pivot. You go from being the best tool for finding sites to being a mediocre tool for analyzing deals. Two completely different products wearing the same logo. I've watched this kill CRE tech companies. They start sharp. One thing, done well. Then a big customer asks for something adjacent. Then another. Each request is reasonable on its own. But after 18 months of saying yes, you're a spreadsheet with a map bolted on. Here's what I told them: Pillar finds you the 3 sites worth underwriting out of 10,000. Then you take those 3 into whatever underwriting tool your team already trusts. We make the funnel. We don't replace the calculator at the end of it. They didn't cancel. But even if they had, I'd have made the same call. The most dangerous feature request in CRE tech isn't the one that sounds crazy. It's the one that sounds completely reasonable.