I keep seeing the same AI adoption gap in CRE: leaders want to start with the hardest work, while their teams are still learning to trust the basics.
‘Show me the one click underwrite!’
You 100% shouldn’t be starting from trying to one shot an IC memo, you have to trust the pieces underneath: the deal screen, the rent roll, the market assumptions. That's how a black box becomes work you can govern.
The performance still comes from your people: the acquisitions lead who deeply knows the nuance of your buy box, the analyst who catches the weird lease clause, the SVP who can say why a deal is wrong even when the model looks fine. AI has to make that judgment faster and more repeatable before anyone will trust it with the bigger calls.
Your deal teams usually ask more practical questions first:
Can it normalise this broker rent roll?
Can it map seller financials into our T-12 format?
Can it pull the lease terms with citations I can defend?
A lot of the market still does not believe the answer is yes. 👀
You see the same pattern inside firms that have spent millions on internal builds. A few power users are doing genuinely impressive work. The long tail is still copying numbers out of PDFs, or waiting for someone else to prove this is real.
The failure mode is simple: companies buy a new way of operating, then try to drop it into the old way of working. That leaves the leaders excited, the users unconvinced, and the expensive AI programme sitting on the side of the real workflow.
Start with one day-one automation your team already understands.
In acquisitions, that might be a deal screen, rent roll normalisation, seller financials / T-12, lease abstraction, or loan abstraction. In asset management, it might be operating statement extraction, a budget variance report, lease obligations, or a debt covenant check. Make the value obvious on day one.
Then make it repeatable: the work has a home, the whole team runs it the same way, and you screen twice as many deals without adding a hire.
Only then do you get to a place where you can take the next steps: inbound deals screened as they arrive, data rooms watched overnight, monthly close started when files land, covenant headroom monitored before breach, IC materials routed with approval and audit. That is where the capital-allocation promise starts to become real.
That is the work: one hated task, made reliable; then one process, made repeatable; then decisions that run on a system the team actually trusts. The firms that are seeing success on their AI adoption journey are the ones that changed HOW work happens, one process at a time.
