My business social media feed has been featuring the same posts for several months, just with different logos. AI-powered deal analysis. Institutional-grade underwriting. One click and you have a full investment report. The screenshots look clean. The copy is confident. The founders have raised real money.
I don’t say any of this as a knock on the people building these tools. Anyone finding product-market fit in a crowded space deserves credit for getting there. But something worth noticing is happening in this category, and it has less to do with the technology than with what the technology is being asked to replace.
The car analogy keeps coming to mind.
When the automobile arrived, hundreds of manufacturers launched. Most of them built functional vehicles. The ones that disappeared weren’t necessarily building inferior products. They were building into a market that was consolidating around distribution, capital, and a few durable advantages that had nothing to do with whether the car drove well. The road to that consolidation was littered with companies that got the product right and the business wrong.
Commercial Real Estate and Business AI underwriting tools are somewhere early on that same road.
The pitch is consistent: stop building models from scratch, stop guessing at comps, stop running formulas you copied from a blog post. Input the property or business financials, get a full analysis in under ten minutes. The efficiency argument is real. Anyone who has spent four hours building a model that could have been scaffolded in forty minutes understands the appeal.
But there is a distinction marketing tends to flatten. Speed and accuracy in calculation are not the same thing as judgment about what to calculate. A tool can run metrics across three scenarios in seconds. It cannot tell you which vacancy assumption is defensible for a specific submarket in the current lending environment, or why your assumptions on a sale might not hold, given where the debt markets are heading. Those answers come from pattern recognition built over years of getting things wrong in real deals, not from a prompt.
What these platforms are genuinely good at is serving people who are earlier in that curve. Someone who would otherwise be guessing gets structure. Someone who would otherwise be slow gets speed. That is real value, and it is a real market.
The question worth sitting with is whether that market overlaps with the one being pitched. Institutional investors and experienced operators tend to do their own diligence on assumptions before they trust a model. They are not looking for a report to hand to a partner. They are looking for someone who can defend the assumptions in the report under pressure. No PDF export solves for that.
The landscape looks like this: a lot of tools competing for the same early-to-mid sophistication buyer, all claiming institutional-grade rigor, all running on largely the same underlying models, none of them especially differentiated from each other on the thing that actually matters to an experienced principal.
That is not a technology problem. It is a positioning problem, and it tends to resolve itself the way these things usually do, which is slowly and then all at once.
The more durable position, for anyone advising clients through deals rather than building SaaS on top of deals, is probably just to understand where these tools earn their keep and where they do not. Used well, they compress the mechanical work. The analytical work, the part where someone sits across from a lender or a seller and defends a number, is still a human problem. It will stay that way longer than the current marketing cycle suggests.
Photo by Birmingham Museums Trust on Unsplash
Leave a Comment