🎧 Check out our latest podcast!

🎧 Listen on Apple Podcasts and Spotify!

One agent fired his $2,000-a-month virtual assistant and handed the job to a $30-a-month bot. He trained it by recording his own screen.

This week on PowerHouseTalk, Tim, Julie, Chris Heller, Dan Lesniak and Tammy Irby dig into what that means for your business: why one human still has to run the AI team, the second brain that keeps AI from guessing, AI assistants that make phone calls now, bold predictions for 2027, what a $1 million sale really pays, and why so many agents are out of practice talking to people.

🤖 A $30 Bot Replaced a $2,000 Assistant

Tammy brought the story that led the show. After a year and a half with a virtual assistant, and a year of saying AI wasn't ready, one agent spent a month testing an agentic AI tool that logs into his apps and does the task end to end. It now writes his show notes, builds thumbnails, schedules his Thursday post, and pulls edited video from Google Drive and uploads it.

Tammy's take: pick one job every week that touches three different apps, record yourself doing it, and hand it over. She's building AI into her processes as fast as she can for one reason: more time in front of her people, the part of the deal AI is never going to replace.

👥 One Human Still Runs the AI Team

Tim's assistant asked him a fair question: once all of this is in place, what am I going to do? His answer: you're going to run it all, because it still needs a human. The Harrises now have 14 AI agents running their two brands, and one person running all of them.

Someone also has to handle the walls AI can't climb yet, like getting logged out, passwords and the sign-ins some tools require. Until that's ironed out, every AI team needs at least one human in the seat.

🧠 Why Your AI Team Needs a Second Brain

Tim's other lesson: the systems, rules and guidelines for your AI team all have to live in one place, a second brain. Without it, the bots slowly lose fragments of what they were supposed to know and make more mistakes. And they rarely admit a mistake. They tell you they hit a roadblock.

Dan added two benefits. A second brain keeps a bot from redoing work it has already done, which saves tokens as AI prices rise. And it cuts down on guessing: Dan has watched AI skip or guess at the numbers in his own workout log instead of checking them.

📞 AI Assistants Make Phone Calls Now

Julie picked the news that Meta's Muse assistant will now place real phone calls for paid users, pitched for the calls you don't want to make: the insurance company, the cable company, the DMV. Her note on that: this does not mean your lead follow-up or prospecting calls. Nice try, though.

Her homework: list your most time-consuming tasks that aren't lead generation, follow-up or presenting. CRM upkeep, support for your transaction coordinator, and setting up showings and tracking down listing agents are all fair game for a bot that calls professionally.

🔮 Bold Predictions for 2027

Dan's top prediction: most real estate predictions will be wrong. Tammy predicts the buyers and sellers on the sidelines finally move. One in five homeowners told Bank of America they plan to buy in the next 12 months, up from 15% last year to 22%. Chris expects a harder market and calls that opportunity: you can't overtake 15 cars in sunny weather, but you can when it's raining.

Julie predicts a big split between agents who make skilled human contact, with AI helping on the important support work, and everyone else. Tim expects a surge of new AI wealth looking for a place to land in luxury, second-home markets and business-friendly states. Get in front of it by networking with luxury agents in those markets.

💸 A $1 Million Sale and the $8,000 Kept

Dan shared a thread where an agent showed the math on a $1 million sale. At 2.5% a side, the agent's $25,000 lost 25% to the brokerage, 35% to a lead referral site, $500 to a transaction coordinator and $1,000 for photos and 3D marketing, leaving about $8,000. The rule of thumb that won the thread: 1% for the brokerage, 1% for marketing, 1% for the agent.

The line every agent needs to hear: a 35% referral fee on a lead you didn't generate is the price of not having a database. Chris asked the other question: are you spending $1,000 on photos for a $2 million home and the same $1,000 for a $180,000 home? AI is already cutting costs like these, from the transaction coordinator fee to software Tammy no longer pays for.

The One Thing to Take Away

AI is freeing up time. The agents who win spend that time on people:

  • Pick one job a week that touches three apps, record it and hand it to AI.

  • Keep one human in charge of your AI team.

  • Give your AI a second brain so it stops guessing.

  • Let a bot make the calls you don't want to make, never your prospecting calls.

  • Build your database so you stop paying 35% for leads.

A lot of agents are out of practice talking to humans. Tammy called past clients this week just to tell them they're on her thankful list and found two who want to move next year. Pick up the phone, and be yourself, because you attract what you are. Optimize for optimism.

- Tim, Julie, Chris, Dan and Tammy
Hosts, PowerHouseTalk

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

Know an agent who should be in this room?
Forward them PowerHouseTalk.

Sponsored by

Get Paid for Data You Already Own

Your business has years of emails, documents, CRM records and call logs. AI researchers need exactly that kind of real-world data, and Mode will pay for it.

You sell a de-identified copy. Your original records and systems remain yours.

  • No broker fees. Mode buys directly.

  • You approve what data gets shared.

  • You never share your passwords.

  • Sensitive information is removed before delivery.

  •  It takes just a few hours of your team’s time.

Depending on data volume and a full review, companies with several years of records could see 100K–5M.

See if your company qualifies and get a quote.