By the time we tie up a deal, Claude already knows it.
That is the part people miss when they think about AI and Investment Committee (IC) memos.
Most teams think about AI as a document generator. Paste in some data, get a first draft. That is useful but it is not the real leverage.
Here is how our workflow actually works:
During underwriting, Claude Code is working inside our acquisition models in real time. It is running sensitivity analyses alongside us. What happens to returns if market rent decreases 10%? If our exit cap is 50 basis points higher? If interest rates go up 25 basis points? It is not just doing the math. It is learning which levers actually move this deal.
By the time we tie it up and need to write the IC memo, Claude has already been inside the deal for days. It knows the assumptions we stress-tested. It knows where the returns are sensitive and where they are not. It knows what kept us up at night.
So when I tell it to draft the memo, I am not asking it to read a model cold. I am asking it to synthesize a process it was part of.
The memo comes back knowing which risks to lead with, which numbers anchor the narrative, and where we built in margin. The human edit is the judgment layer: why this market, why now, why us. The rest is compression.
Our last IC memo took half the time and it was better.
If you are using AI only at the end of your workflow, you are leaving the best part on the table.
What other parts of buying and operating commercial real estate have you found most useful to bring AI into?
The Best AI Work Happens Before the Memo
