SF Bay Area · Member since 2020 · 10 posts · 6 votes
Hi, I am interested in connecting with real estate investment and development companies looking to accelerate due diligence and improve compliance with AI. More specifically, analyzing hundreds of pages of text in disclosure packets including inspections and advisories to understand the risks when closing deals.
AI's decent at flagging missing docs or inconsistent numbers across a rent roll, but it can't judge intent, and that's where liability actually lives. We handle due diligence support across a bunch of states and the pattern's consistent: automation speeds the first pass, but someone still owns the sign-off. Where it earns its keep is triage, sorting hundreds of pages down to the ones needing real eyes. Where it falls short is title exceptions, zoning ambiguity, anything needing judgment.
Compliance-wise, keeping an in-house team current on every jurisdiction gets expensive fast. Salaries, training, turnover add up quicker than people expect. Glad to compare notes on how outsourced ops handle that if it's useful.
It is interesting to see this post resurface from 2 yrs ago, and scary how fast time goes by, given it doesn't feel that long ago that the conversations just started about "how can we better use AI".
I am going to mirror Keith's comments: reviewing DD data dumps for missing documents, or extraneous documents certainly has helped me. Also, the initial review of broker packages has been very helpful, with AI researching local market conditions, LoopNet and Crexi details for comparable sales and leasing rates, news stories about new developments coming into the market. I would say it is great at flagging deals to kill.
And it is great to run my detailed underwriting against to try to catch issues and general reality checks on whether I typoed something, double entered an income line item, as well as at least calling out possible issues with assumptions.
But it all comes back to me. I find it can be overly conservative at times, which is better than the alternative. It cannot deal with ambiguity, which is nice to have it at least flag where ambiguity lies, etc.
This is an interesting use case for AI. I work in commercial real estate lending, and due diligence is definitely an area where there’s an enormous amount of information to review and identify potential issues early.
I’d be interested in learning more about what you’ve built, particularly how you’re using AI to identify and prioritize risks within inspection reports, environmental reports, disclosures, and other due diligence documents.