Transcript
Have you noticed that most real estate investors are asking AI the wrong question? They open ChatGPT or Claude, they put in some information about a deal, and then they ask, "Should I buy this?" And I think that's exactly the wrong way to use AI. AI should not be making investment decisions for you. But it's surprisingly useful at something else. Trying to kill your deal.
And I think that's a much smarter way for real estate investors to use it. When we look at a multi-family deal, we're obviously trying to figure out whether the investment works, right? But there's a danger here. The more you like a deal, the easier it becomes to defend your own assumptions. Maybe you think rents will grow 3%.
Maybe you think insurance is going to stabilize. Maybe you're assuming assuming that you'll refinance in 3 years. And maybe all of these assumptions are perfectly reasonable. What I want is somebody sitting across the table saying, "Wait a minute. What if you're wrong?" And that's where AI gets really, really interesting.
So, instead of asking AI to confirm that this is a good investment, I actually give it the opposite job. I tell it, "Your job is to convince me not to buy this deal." And then suddenly you're having a really different, really interesting conversation. Because now AI isn't trying to help you justify the investment. It's looking for holes in it. Right?
For example, maybe I've assumed rents are going to grow 3% every year. Okay, what happens if they don't, right? Maybe maybe we've also assumed insurance costs are going to increase at a certain rate. What if they're significantly higher? Interesting questions.
Maybe we're planning to refinance. What happens if interest rates don't cooperate? Maybe our exit cap rate is too aggressive. There's a lot of interesting stuff. Maybe there's a huge amount of new apartment supply that's coming into the market that we haven't properly accounted for.
Or maybe the biggest problem is something that we haven't even thought to ask about. So, that's what I want AI looking for, right? Not reasons to say yes, but reasons to say no. And you can make this really really systematic, right? Here's essentially what I ask it to do, right?
Find the five biggest risks. Tell me which assumptions look the most optimistic. Tell me what information's missing. And then stress test the deals so that the rents don't grow, expenses increase, or the cap rate go up. You know, I want the information, right?
And then give me the strongest possible argument for walking away. Right? Think about what we're doing here, right? We're deliberately trying to break our own investment thesis before we put money into the deal, and that's incredibly valuable because finding a bad assumption before you close, you know, costs almost nothing. Finding it after you close could cost millions.
Now, there's one important warning here. AI can be wrong, it can make up stuff. So, I specifically tell it, "Separate verified facts from assumptions, and do not invent missing numbers or marking data." Important, super important. If AI doesn't know something, I want it to tell me that the information is missing. I don't want it to fill in the blank with some number it found or even invented.
And then there's another mistake I don't want you to make. Don't ever let AI make the final investment decision. That's our job. AI doesn't replace underwriting. It doesn't replace market knowledge.
Uh it doesn't replace due diligence. It certainly doesn't replace experienced people, right? So, think of it as one more person, really important person, but one more person sitting at your investment committee table. Except this person's entire job is to be annoying. Is to challenge you, to question your assumptions, to point out things that you may not want to hear.
Now, all of that's useful, right? So, if you want to try this on your next deal, here's the prompt. Act as a skeptical institutional real estate investment committee. Your job is not to help me justify this multifamily deal. Your job is to find the strongest reasons that I should not invest.
Identify the five biggest risks. Identify assumptions that appear too optimistic. Tell me what information's missing before I make a decision. Stress test what happens if rents don't grow, expenses rise, or exit cap rates increase. Give me the strong strongest argument for rejecting the deal.
Separate verified data from assumptions. Do not invest missing numbers or market data. And then recommend one of four things. Proceed. Proceed with caution.
Do more research. Or pass. And of course, explain your recommendations in plain English. That's it. You don't need some complicated AI system.
You don't need to build software. You don't need AI to predict the future. You're simply giving it a job that it's actually pretty useful at, you know. Challenge thinking, find blind spots, and tell me what could go wrong before my money is on the line. I think that's one of the smartest ways real estate investors can use AI today.
Not as a crystal ball, but as a skeptic that's sitting at the investment committee table. If you'd like more tools, webinars, and resources to help you become a better real estate investor, please join our Investor Club. It's completely free. Go to multifamilyu.com/club. Thanks for watching, and I'll see you in the next video.
Bye-bye.