AI is not as effective as you think it is

Real Estate Technology
3wFollow

AI is not as effective as you think it is

Adam MaciasPro Member
Real Estate Consultant · Fort Collins, CO · Member since 2022 · 284 posts · 178 votes

The biggest flaw of AI is like a  salesman paradox. Just like with salesman, they become a little sociopathic over time because when you’re commission based only, your emotions get so suppressed that you forget what it’s like when someone tries to sell you something. You start shifting your mind into a “If they say no it’s their loss” but the truth is some people truly don’t need your product, don’t care for it, or your product just may suck if you really took time to assess why no one wants to buy it or if they keep giving negative reviews. But salesman don’t see it that way. They see it as a “okay NEXT” way of life. And it’s a terrible way of life. There’s more to life than constantly living off commissions. Eventually, you need passive income and time freedom. Which most commissioned roles like being an agents or sales rep don’t give you for a while. Which is why I see the attraction towards utilizing AI to help you “automate” and make more money. However, if you look at the data of how many people in the United States dislike and outright hate AI use, it should make you reconsider what you use AI for. “Virtually all Americans dislike robocalls, with surveys showing that an overwhelming majority—up to 94% of U.S. adults—find them annoying and disruptive.”

This took a simple Google search to find and look into surveys about this. If you hate robocalls and AI calls, what makes you think your prospect will like them? Just because you have something to sell, doesn’t mean anyone is obligated to buy. There’s a few gurus I see selling courses and systems on automating your calls. Huge red flag already given practically everyone hates getting a robo call but agents and investors gobble it up and think they’re being productive. I’ve yet to truly find anyone prove to me that because of utilizing AI agents for their calls, it actually made them more money. Heck, even an investor I spoke with last week said he’s making thousands of calls a week to cold call homeowners to get deals for his wholesale business. If you have to make thousands of calls, that doesn’t sound very effective to me and it sounds like a pending lawsuit eventually when a litigator catches you. There are laws against using automations for cold calls.

But people are going to do what they’re going to do. I just advise not to buy into the hype of using this for any of your prospecting. That wholesaler making thousands of calls a week still hasn’t gotten a deal this month just FYI. Because if you do what everyone else is doing, then you no longer stand out. This is the AI paradox happening just like the salesman paradox. You think more means more effective. But really, it’s just more busy work.

Disclaimer: Entertainment purposes only. Not legal advice.

(lol)

0Reply
182 views

2 Replies

Jump to latestLatest
  • Michael Eskenasy
    Investor · Pacific Northwest · Member since 2026 · 510 posts · 285 votes
    3w

    I think you’re seeing a real failure, but you’re seeing it at the point where it finally becomes obvious.

    The robocall is the crash. The problem started much earlier.

    What’s happening underneath a lot of these AI systems is that they’re being asked to reconstruct reality over and over again from whatever information happens to be available at that moment. A CRM record, an old email, a note somebody entered six months ago, a new instruction buried in a thread, a changed circumstance nobody updated, a conversation another person had and never logged. The model looks at all of it and tries to figure out what probably matters now.

    That can work remarkably well. Until it doesn’t.

    And when it doesn’t, the failure looks stupid. Someone gets called who already said no. A person gets the same outreach twice. A prospect is treated like a stranger after six months of conversation. The system says something technically reasonable that is completely wrong for that particular human being.

    That’s the part people experience as “AI sucks.”

    But the deeper issue is that most AI was not designed to maintain the living continuity of a human relationship or a business over time. It was designed to take information in, find patterns, and generate a useful result. That is an extraordinary capability, but it is not the same thing as preserving reality.

    A business cannot operate forever on “what probably happened.”

    At some point it has to know what happened.

    It has to know that this person said not to call again. That the owner changed the price yesterday. That the lender already rejected this structure. That somebody promised a follow-up on October 1. That a newer instruction overrides an older one. That an email was sent. That a document was received. That something is still unresolved. That one source is authoritative when three systems disagree.

    Those sound like mundane details. They are not.

    They are the connective tissue that makes an organization appear coherent to another human being.

    Without that connective tissue, every AI interaction is a tiny reconstruction exercise. And every reconstruction introduces the possibility of drift. Usually the drift is small enough that nobody notices. A detail gets slightly distorted. A previous decision loses its reason. An old fact survives longer than it should. One system updates and another doesn’t.

    Then it happens again.

    And again.

    Nothing looks catastrophic. The machine keeps working. The dashboards are green. The automations fire. The emails go out.

    But underneath, the organization is slowly moving away from what is actually true.

    That’s Black Ice.

    You don’t notice the loss of traction while it’s happening. You notice it when the car is already sideways.

    That is why I don’t think the answer to bad AI prospecting is simply “use less AI,” and I definitely don’t think the answer is “AI isn’t effective.”

    The real mistake is treating the model itself like the operating system.

    We learned the hard way that it needs infrastructure around it.

    Something has to preserve the facts that should not be reinterpreted every time. Something has to know which instruction is current. Something has to remember what already happened so it does not happen twice. Something has to reconcile contradictions instead of casually choosing whichever record happened to surface first. Something has to keep track of unfinished obligations. Something has to notice when the world changed after the last decision was made.

    And most importantly, something has to sit between what the AI proposes and what actually happens to another human being.

    I think of that piece as a shock absorber.

    The model should be allowed to move quickly. It should be allowed to reason, search, synthesize, challenge assumptions, and come up with a better answer than the one we had yesterday.

    But movement inside the model should not automatically become movement in the real world.

    Before an action reaches a person, the system has to ask a few very human questions.

    What do we already know about this person? What happened before? What did we promise them? Has anything changed? Are we sure this hasn’t already been done? Is there a newer instruction? If two records disagree, which one should we trust? If we do this now, what does the next person need to know tomorrow?

    That is the difference between using AI as a tool and building what I would call human cognitive infrastructure around it.

    The purpose is not to make the machine more human.

    The purpose is to keep the human world intact while the machine moves through it.

    That distinction matters because people experience continuity, not databases.

    They remember the conversation from last week. They remember what you promised. They remember that they told you their father was sick. They remember that they already explained why they cannot sell right now.

    If your system cannot carry that forward, it does not matter how natural the voice sounds.

    It will still feel fake.

    And once you see that, the thousands-of-calls example becomes almost beside the point.

    Of course that eventually fails.

    You took a technology that is excellent at generating and scaling output and connected it directly to human beings without first solving continuity, memory, authority, context, and restraint.

    You did not automate intelligence.

    You automated drift.

    That’s the part of this conversation I think we are still missing.

    0Reply
  • Richard F.
    Honolulu, HI · Member since 2008 · 3k+ posts · 2k+ votes
    3w

    Or, it could just be that there is a LOT more to this "REI" than the average person thinks. A lot more options. A lot more solutions. A lot more variables. A lot more "definitions". A lot more "Unknowns". Which all result in a lot more surprises...usually at your expense.

    0Reply
Join the conversationCreate a free account to reply, vote on answers and follow this thread.
Log InJoin for Free