🎧 Listen on Apple Podcasts and Spotify!

One agent. A $5,000 AI assistant. 60,000 old leads. And a reported 34 closings and $490,000 in commission.

It is exactly the kind of story that spreads through agent groups overnight. This week, with most of the panel at EXPCON in Salt Lake City, Tim, Julie, and Chris took it apart number by number.

Their verdict: the funnel probably doesn't add up, and the way it was reportedly worked could expose an agent — and their broker — to serious legal risk. But there is still a useful idea buried inside it.

This week's PowerHouse Talk also covers AI coaching and role-play, cinematic listing films, AI agents that make phone calls for you, and listener questions about AI transaction coordinators, staging costs, and a $1,900 credit that never made it to the closing statement.

Julie's win this week came from her coaching calls: many of her clients are on track, or a little ahead of their goals, in what she called a very challenging sales environment. Rates are higher. Prices aren't coming down.

So why are they ahead? According to Julie, they are:

Tim asked the follow-up question every agent in that position should hear: what is the greatest risk to an agent riding the wave of their best year?

The answer: with a pipeline full of contracts, the first thing to go is lead follow-up, marketing, and prospecting. Over-celebrate the fourth quarter and you set yourself up for a thin first quarter.

Chris shared his own win: a past client from roughly 20 years ago recently joined his team — and just landed her first appointment.

Tim raised a question that has bothered him for years. Why do agents build elaborate systems, 77-touch campaigns, and complicated direct-mail plans when, as he put it, the vast majority of business comes from centers of influence and past clients?

The panel offered a few answers:

Julie's coaching fix is to break the work into micro, bite-size pieces. The agent who has turned calling expireds into a huge, scary project starts with three to five contacts a day. The script doesn't have to be perfect. The work just has to get done.

And within a week or two, she said, they almost always say the same thing: "I can't believe how easy that was."

Chris then reframed the whole AI conversation in one line:

Tim added that AI will replace the version of you that was doing the non-human work. The human work — reading people, building relationships, meeting more people — is where agents need to get really good.

The story, originally posted on Reddit, claims one agent spent $5,000 on an AI bot that texted and called a database of 60,000 dead leads. The reported funnel:

Texting and calling fees reportedly added another $6,000. The poster's explanation for why it worked: "People relax when they're talking to something that's simply checking in."

The panel wasn't buying most of it. The numbers don't hold together — 60,000 leads minus 11,000 scrubbed numbers isn't 44,000 contacts. Other versions of the story say the outreach was done by email rather than text, and sending that much email would require bulk senders and rotating IP addresses, with most of it likely landing in spam. And few agents, if any, really have 60,000 leads.

Fact or fiction, the bigger issue is what would happen to an agent who tried to copy it.

The first red flag for the panel wasn't the math. It was the texting.

If those consumers never opted in, Chris said, an agent doesn't have the right TCPA approvals. Add numbers on the Do Not Call list and you are walking straight into a lawsuit. And it may not stop with the agent — the broker can be pulled in, too.

The panel also highlighted a risk most agents never think about: reassigned numbers. Even if you had a legitimate business conversation with someone six months ago, if that number has since been given to someone else, your text may now be a violation. Multiply that across thousands of old leads and the exposure adds up fast.

The panel's other cautions:

Chris made it personal. About three weeks ago, he was served over text messages sent to a neighborhood in 2024 — community information with a landing page that featured his listings. The texts didn't go out from his team, and he never paid for or agreed to the service. His advice: it's not worth the risk.

And if you ever do end up in litigation (not legal advice): get a really good attorney, never try to hide or delete anything, and understand that many of these firms are settlement machines that like to settle early.

Julie steel-manned the story. Most of it may be bunk, but the underlying idea is worth testing.

Every agent has a pile of old leads they stopped calling. Instead of 60,000, think about a couple hundred you've put on ice — old open-house leads or people you haven't talked to in a while, with notes in your CRM. Have an AI send a simple message, perhaps by email, not in huge bulk, and test it.

Tim explained why this is different from the old CRM drip. Canned emails gave everyone the exact same experience. A well-built AI assistant can actually converse: answer questions, follow up, ask its own questions, and engage the way a human would.

His rule for doing it right: disclose that it's AI. It's the ethical thing to do — and he believes most consumers won't care, as long as the experience is helpful. As he put it, this is the worst the technology will ever be.

Tim shared what he has spent the last eight months building into the new Premier Coaching:

Chris's reaction was simple: it only benefits you if you use it. Otherwise it's like attending a seminar, taking great notes, and never looking at them again.

Tim made a bold prediction: the days of the standard listing video — the flyover, the zoom, the drone footage — are numbered, especially in the upper end.

His example was a horse property in Dallas. Instead of photos and a drone pass over the barn, imagine a short cinematic film:

Two years ago, that would have meant storyboards, actors, a film crew, and a budget almost no listing could justify. Now, Tim said, you can load still photos or iPhone clips into an AI video engine, prompt the scenes you want, and produce it — and for anything around three minutes or less, the cost is minimal.

He went a step further: sellers could click optional prompts on a listing page — a family at Christmas, the stars at night over the property — and watch a new scene render.

Chris added that agents who don't want to learn the tools can pay a company to do it. He also shared how a friend made AI videos of their group at a music festival in Austin — hilarious, and a little scary.

On where the line is, the panel agreed: using the property's actual photos and showing what a 12-acre horse property looks like at Christmas with snow isn't going too far. Adding something that isn't there is.

Chris has spent the last few weeks using a personal AI agent with its own phone and computer. It calls businesses on his behalf — booking the restaurant that doesn't take online reservations, changing dentist appointments, fighting billing disputes.

What impressed him most is that it's proactive. It noticed he had two flights booked on the same day and asked whether to cancel one. It flags delays and offers alternatives. It drafts replies and asks whether to send or edit them.

His admin's review after a week: "By far the best AI tool I have used."

Tim called out the shift that matters: AI isn't just finding the problem and handing you a list. It finds the problem, proposes the fix, and deploys it.

The flip side is coming fast, too. Buyers' AI agents will soon be calling your office to book showings. Is your team ready to talk to a robot?

One of this week's headlines answered part of the question: on one tough benchmark, the latest Claude Sonnet model jumped from roughly 10% to 70% in a single release. What frustrated you in the spring may work now.

Julie's answer: yes, give it another shot — but real estate is not set it and forget it. You are the one with the license, and you still have to oversee deadlines and contingencies.

Tim described how the AI team inside the new Premier Coaching handles that: an AI transaction coordinator manages the dates and tasks, and a lead AI team member, Avery, checks its work — so there are two layers of review before anything reaches the agent.

Julie said it depends on the market and the price range. Many of her coaching clients, up to roughly the $2 million to $3 million range, skip full staging and pay for a two- or three-hour consultation instead. The stager helps the seller declutter, neutralize, and keep only the key pieces that look good — and nine times out of ten, that does the trick.

Some agents keep their own curated collection of décor in storage so they don't have to write big staging checks.

If paying for staging makes business sense, the panel's advice is to have the seller pay the stager and reimburse the seller at closing. That way you aren't writing checks before you know the deal will close.

Chris was direct: your TC works for you. Your TC made the mistake. Own it and make it right. "It's not that complicated."

The rest of the panel added that the real concern behind the question is legitimate — moving money around after closing and outside escrow can create a bigger problem. Fix it, and fix it the right way.

The $490K story will keep circulating. Some agents will try to copy it. Some of them will end up talking to attorneys.

The better lesson is this:

The technology isn't going to leave anyone behind. The lack of human skills will.

— Tim, Julie and Chris
Hosts, Power House Talk

whylibertas.com/harris
whylibertas.com/heller

💬 Want your question featured on an upcoming episode? Reply to this email or DM us on Instagram— we might break it down next.

📩 Know an agent who should be in this room?
Forward them Power House Talk.