Every operator who has been in this business a while is sitting on an asset they have never put to work: their own email archive.
Think about what twenty or thirty years of deals actually leaves behind. Every LOI you sent. Every deal you underwrote and the ones you passed on. Every buyer who told you what they wanted and every seller who told you what they would take. All of it is sitting in your inbox, unstructured and unsearched, which is exactly why almost nobody treats it as data. I think that is about to be one of the most valuable things a lot of CRE professionals own.
Why the edge is moving to data
Analysis is getting commoditized fast. The same models that read an OM for me read it for the broker across the street, at the same quality, for the same twenty dollars a month. As the tools get more capable, the gap between a good analyst and a great one narrows, because the model floors everyone at a high level. When the analysis is a commodity, the edge has to come from somewhere the model cannot reach. That somewhere is your proprietary data.
The base model knows the public internet. It does not know what a specific buyer offered on a deal that never closed in 2019, or which lender actually funded when the market got tight, or what a comparable building is really getting in rent net of concessions. You know those things, or rather, your email knows them. That information is genuinely yours, and it is the part of the business that does not get easier to copy as the models improve.
Turning the archive into something you can query
The practical move is to stop treating your email as a place you search one message at a time and start treating it as a dataset.
A simple version: point an AI tool at a defined slice of your inbox, say every deal email from a given market over the last three years, and have it pull the recurring facts into a structured table.
Asking rents and what things actually traded at. Who bid and at what number. Which assumptions kept showing up in your own underwriting. What you end up with is a private comp set and a buyer-preference map built entirely from your own history, the kind of thing a CoStar pull will never give you because it never happened in public.
You can run this on a schedule instead of one-off. A recurring job that reads new deal emails as they arrive, extracts the same fields, and appends them to your dataset, so the asset compounds while you sleep. The first pass is the hard part. After that you are maintaining a living record of everything that crosses your desk.
One caveat, because this matters: the output is only as clean as the inputs and the rules you give it. An AI reading your email will happily invent structure where there is noise. You have to define the fields you care about, spot-check the early extractions, and treat the first version as a draft an analyst handed you, not gospel. I have watched these extractions look perfect and be quietly wrong on the numbers that matter most.
What this looks like at scale
The operators moving fastest on this are not waiting for a vendor to sell them a product. One group running several thousand units built their own internal command center, a place where the firm’s own data is queryable instead of buried across inboxes and spreadsheets. It took months and it was not free. What they got was a tool no competitor can buy, because it runs on data no competitor has.
Starting does not require a command center. Pick one question your inbox can answer that public data cannot, and build the smallest version of it.
The part people underrate
As information becomes cheaper to get, the scarce thing becomes the proprietary record of relationships and decisions that only you have, plus the discipline to turn it into something you can use. The firms that win the next few years will be the ones that realized their most valuable dataset was the one they had been letting rot in an inbox. Your email has been collecting it the whole time. Pick one question and start there.
We break down the exact workflows for this, turning your own email and deal history into a private comp set, inside the AI for CRE Collective. If that is useful, come find us.
