Hey guys, long time no talk.

Sorry for the slower pace of real estate content on Twitter & the newsletter.

I’ve been working really hard with my cofounder Arjun on our AI app, but I also wanted to think about what the future of real estate investment management looks like and how we can implement AI into it.

The State Of AI In CRE

AI is obviously here to stay. I think the hype of “AI taking over our jobs” is overhyped in it’s current form, but there are plenty of useful applications where it can augment manual work flows.

For example, I do not think that AI cannot replace human analysts just yet. I think they can help streamline the information gathering & analysis process, but can’t do a full underwriting just yet.

In the future, I think it definitely is inching towards that. I like to think of where workflows are going to be in 3-5 years. The more inputs & training data is available, the better AI can synthesize a decision based on its training data.

Real Estate’s Big Aversion To Sharing Data

Historically, Costar is the big dawg in the space. Corelogic, Appfolio, Yardi, Realpage are some of the other giants in the space. Compstak (my alma mater) is also in the running for CRE leasing/sales data.

There are 13 nondisclosure states in the United States, examples being Texas & Utah. This makes it where sellers & buyers don’t have to disclose sales prices, buyers/sellers commissions…

It’s great for people who value privacy and don’t want public records of what they paid for property. But it means one problem.

The trouble in CRE is the fact that a lot of the data is either siloed, proprietary, or heavily guarded. This makes it unconducive to LLMs training on publicly available data.

This means the big dawgs who have gathered a huge data bank or pay up the nose for an API will have access to data where they can train insane LLM’s. I imagine that would be worth a pretty penny.

Imagine if Costar was able to tell you what specific stats about a building’s financials via chatbot or you could tell it to create a pro forma with X, Y, Z assumptions.

Where Can AI Be Helpful

Market Research and Analysis

AI tools can generate comprehensive market reports by analyzing data from various sources, including news articles, market reports, and industry publications.

A lot of brokers/investors are already doing this right now. They use ChatGPT/Claude for property descriptions & skimming market reports to get key insights.

You can also use AI to extract key data & phrases to get an overall view of the market. It’s a task that used to take you 30-45 minutes to read a report cover to cover that returns information within seconds of entering those keystrokes.

Our attention spans are cooked.

Lease Administration

There are firms like LeaseLens and Prophia. They’re tools used to allegedly help get lease abstracts for dense commercial real estate leases into plain English. Save the JD for the legalese.

With multifamily, it’s pretty easy to understand what traditional leases are like. I imagine the base version of Claude/Gemini/ChatGPT can do the same thing too.

Predictive Maintenance

This might be a shot in the future, but I anticipate more buildings are going to be using IoT (internet of things) infrastructure. IoT means a lot more sensors + data. If it’s open source, this means that there’s a lot of potential for AI to get better data about buildings and when things need to be replaced.

AI Underwriting

This is a problem I’m working on… AI is great to parse data from things that were very manual previously & you can create financials on it.

Enter DealIntel

Dealintel.ai is a SaaS app I’m working on.

The flow looks like this:

1. You upload deal docs (OM/Rent Roll/T-12) in Excel or PDF form.

2. AI will pull all the data out include financial information, unit mix

3. You can export the data and reference those cells in your model.

4. We can create a rudimentary pro forma based on the financials as well.

Essentially, the benefits are:

• Analyzing more deals

• Automates parsing of data and exporting

• Reduces amount of time needed for quick decisions

We’re looking for initial users. If you work in institutional real estate private equity in multifamily or lend on multifamily, please respond to this email.

Thank you for reading.

-Mo