The property condition assessment came back at 94 pages.
It cost $6,400 and took three weeks. It contains, somewhere, the two facts that will determine whether this acquisition works: the remaining useful life of the roof, and whether the electrical service can carry the load the new use requires.
Two facts. Ninety-four pages. And the person who needs them is going to skim, find something that looks like an answer, and move on — because he has four other deals and it's Thursday.
This is the actual condition of information in a real estate business. You commission expensive, rigorous documents, and then you underuse them, and the gap between what you paid for and what you extracted is the most quietly wasteful thing in the whole industry.
The problem is not that reports are long
Reports are long because the world is complicated and the engineer is being careful. That's correct. You want the appraiser to document his comparables and his adjustments. You want the environmental consultant to note the dry cleaner that operated on the adjacent parcel in 1987. The length is a feature.
The problem is that documents are built to be read linearly and used interrogatively. Nobody, ever, sits down and reads a Phase I from page one. What they do is arrive with a question — is there a REC on this site? — and go hunting for the answer.
So the format and the use are mismatched, and the tax on that mismatch is paid in hours, in things missed, and occasionally in a very expensive surprise nine months after closing.
What "asking the document" actually means
You take the whole corpus for a deal — the PCA, the appraisal, the Phase I, the title commitment and exceptions, the survey, the leases, the operating statements — and you put it into a system that can read all of it and answer questions with citations.
Then the work changes shape. You ask: what's the remaining useful life of the roof and where does it say that? And you get: 5-7 years, PCA page 41, with the paragraph. You ask: list every capital item the PCA flags with a cost above $25,000, with page references. And you get a table, in ten seconds, that would have taken an analyst two hours to assemble and would have missed one.
You ask the questions across documents, which is where it gets genuinely valuable: does the appraiser's assumed occupancy match the actual rent roll? Does the survey show the easement described in title exception 11? The seller said the HVAC was replaced in 2021 — does the PCA agree?
Those cross-document questions are where deals go wrong. They are also exactly the questions nobody has time to ask, because asking them manually means reading two long documents simultaneously and holding both in your head.
Where it must not be trusted
Here's the discipline, and it's non-negotiable.
Every answer must cite its source, and the human must be able to click through to the actual paragraph in the actual document in one motion. An uncited answer is worthless — worse than worthless, because it's confident.
You do not make decisions on the summary. You use the summary to find the page, and you read the page. The system's job is retrieval and synthesis, and the human's job is judgment on the primary source. That distinction sounds pedantic until the first time a machine confidently paraphrases a qualifier out of existence, and you're the one who signed.
And nothing about this replaces the consultant. The PCA is still worth $6,400. The engineer still walks the roof. What changes is that you now actually consume what you bought.
The number, and it isn't the obvious one
You could price this in hours. An analyst spending eight hours per deal on document review, across twenty deals a year, at $50 loaded — that's $8,000. Real, but not thrilling.
The number that matters is the one nobody puts on a spreadsheet: what does it cost you to miss something?
One easement misread. One deferred maintenance item that didn't make it into the underwriting. One lease with a termination right nobody flagged. Any single one of those, once, dwarfs a decade of the labor savings. And the honest truth is that these things happen at every shop, periodically, and get absorbed as the cost of doing business.
They're not the cost of doing business. They're the cost of reading 94 pages on a Thursday when you have four other deals.
What your people do instead
The analyst stops being a search function and starts being an analyst.
The acquisitions lead walks into the investment committee having actually interrogated the document set rather than skimmed it, and can answer the hard question from the partner who's been doing this for thirty years — because the partner will ask about the easement, and the right answer is the paragraph, not a shrug.
And the deal team gets faster in the way that compounds: not because anyone is rushing, but because the retrieval is free, so the thinking gets the time it deserved all along.
You already paid for the information. You paid a lot for it, and you paid for it precisely because it might contain something that changes your mind.
The least you can do is ask it.