🏠 51% of Sellers Check Your Listing Price With AI - Here's What to Do

UPGRADE WITH LIBERTAS & EXP REALTY

🎧 Listen on Apple Podcasts and Spotify!

You gave them your number. They nodded politely. Said they wanted to sleep on it. Walked you to the door. Twenty minutes later, they're at the kitchen table typing your number into their phone.

A recent survey of 1,000 sellers found that 51% ran their agent's price recommendation through AI, and 43% used it to second-guess other advice they'd been given by their agent. Not one of them brought it up on the follow-up call.

They just came back a little cooler on your number — or they didn't come back at all, and you never found out why. That's the listing appointment now. You can't win this by arguing with the AI.

You can't win by trying to convince a fact-checking seller that ChatGPT is wrong. You win by giving the seller the right prompt to use, before they invent a leading one on their own.

Today we walk through the specific pre-qualification adjustment, the exact ingredients of a professional pricing prompt, why the prompt controls the answer, the two-point financing gap that's silently killing your resale listings, and the AI-first playbook that turns your fact-checking seller into a client who trusts your number.

You now have two pricing conversations to manage. Not one.

The old Zestimate problem never went away. Sellers still walk into listing appointments with an inflated number in their head from their favorite home-value site. Your competitor is still the agent who overprices to buy the listing, then chases it down with price reductions three months later.

The new AI problem stacks on top. After you leave the appointment, they open ChatGPT, Claude, or Gemini and type in some version of "Is $525K the right price for my house?" — and the AI, being a people-pleaser designed to match tone, tells them what the way they asked implies they want to hear.

Which means the seller now has three conflicting numbers:

And your number is the fourth. Sitting in the middle. Now competing with three data sources the seller trusts more than they should — and than they trust you.

Do not argue with the AI. Do not try to prove ChatGPT wrong to the seller. Do not go into the follow-up conversation ready to correct their assumptions.

You will lose that fight 100% of the time. The AI is confident, articulate, and available at 11pm when you're not. Your seller will side with the confident, articulate 11pm friend over the human who's asking them to accept a lower number.

The winning move is inverted: instead of fighting the AI they'll use, give them the AI prompt yourself, before the appointment. Now you're both running off the same instructions. Now the answer they get isn't biased by how they happened to phrase the question at midnight. Now your number and the AI's number are usually within a few percent of each other — because both are running on the same objective input set.

Before the listing appointment, add one question to your existing seller pre-qualification:

"When you've been doing your own research on the home, what AI tools have you been using? ChatGPT? Claude? Gemini?"

Almost every seller will name one. A few will deny using any and then quietly use one that night. The point isn't to catch them. The point is to know which platform they're likely to check you on so you can go do the same check yourself, first.

Before you present your number at the listing appointment, do this:

Open the same AI the seller will use. Ask it — as if you were the seller — to price the home. Use whatever leading, biased, emotionally-loaded phrasing a nervous seller with a mortgage decision on the line might use. "What should I list my 3-bedroom in [neighborhood] for? I've put a lot of work into it and I don't want to leave money on the table."

Note the number the AI returns. That's the number your seller is about to see. Now you know exactly what you're walking into.

If the AI's number is close to yours — great, the appointment is going to be easy.

If the AI's number is 8-15% higher than yours — which is common — you now have a decision to make about how to handle the gap professionally, in advance, rather than being blindsided by it on the follow-up call.

Modern AI is designed to be conversational and helpful. That design has a specific side effect: it matches the tone and framing of the person asking.

Same house. Same AI. Three completely different numbers. The difference is the prompt.

Your seller doesn't know this. They think AI is a neutral oracle. In fact it's a mirror that shows them what they wanted to see, unless the prompt forces it not to.

The prompt you give your seller should require the AI to:

Here's the shortcut. Don't try to become a prompt engineer.

Ask AI to write the prompt for you.

"I'm a real estate agent. I need a prompt I can share with my sellers before their listing appointment. The prompt should have them run a rigorous CMA on their own property that avoids the biases most sellers accidentally introduce. It should require sold comparables only, factor in new-construction competition and builder financing, use current interest-rate environment, and account for current local days on market. Write me the prompt in language a homeowner will actually copy and paste."

Ten seconds later, you have a professional-grade pricing prompt. Save it as a template. Send it to every seller before every listing appointment.

Before deploying the prompt to sellers, run it yourself on a property you already priced. Compare the AI's output to your own CMA on the same house.

Ship a prompt you've validated, not one you drafted and hoped worked.

Important caveat. AI answers with total confidence, even when it's substantially wrong. It will not say "I'm not sure — you should verify this with a local expert." It will give a number, in a well-organized paragraph, with plausible reasoning, and that number can still be off by 20-30%.

Real example: Recent South Florida transaction. A buyer had ChatGPT tell them a $13M property was worth $10M. The buyer refused to go above $10M. AI was working from public data only. It hadn't seen three off-market sales in the previous six months that would have justified the $13M ask.

Result: the buyer either walked or the seller took a huge haircut. AI's confidence made the outcome worse, not better.

Related tactic. When your AI-run CMA differs from the seller's AI-run CMA, don't try to prove the seller wrong. Show them what your AI knew that theirs didn't.

What professional agents can feed their AI that a homeowner typically can't:

Show the seller the difference. "Here's what your CMA saw. Here's what mine saw. Here are the three sales you couldn't have known about that change the number." You just proved you're doing something the AI can't — and the seller respects you more for it.

Zoom out for the strategic point.

In most current US markets, the pricing question isn't really about price. It's about monthly payment.

Almost every buyer at every price point below ultra-luxury shops on payment, not price. Same reason nobody balks at the total cost of college, a car, or an iPhone financed over 36 months. What people react to is the number they see on the monthly statement.

Which means: when your seller's home is competing with new construction, the price comparison isn't $450K vs. $600K. It's payment vs. payment.

Here's the current number that most sellers don't know and most agents underestimate:

As of this week's recording:

That's roughly a two-point spread.

And it gets sharper. Some large builders are running first-year buydowns at approximately 1.875% — start rates that dramatically alter the monthly payment for at least the first 12 months.

Practical implication: on a payment basis, a $450K resale at 6.75% often competes head-to-head with a $600K new-construction house at 4.75%. If you show a buyer the resale and the new construction side by side with the same monthly payment, the resale loses much of the time — because the buyer gets more house for the same monthly outlay.

This is happening in a lot of markets right now, and your sellers don't know. Neither does their AI, unless you tell it to look.

Related. Most new-construction inventory isn't in the MLS. Which means the seller's AI CMA, by default, is missing most of the actual competition.

Explicitly instruct any AI CMA — yours or the one you send the seller — to search:

AI won't do this by default. You have to tell it to.

Now the payoff moment. The seller comes back with their fact-check number. You come back with:

Their CMA looks like homework. Yours looks like a professional pricing engine. In almost every case, the seller now trusts your number — not because you argued them into it, but because you demonstrated a level of research their AI wasn't asked to do.

None of your competitors are doing this. Most of them are still fighting the AI. You're using it better than the seller.

Related tactical point. Regardless of what rates do next, there are thousands of listings in every market currently expiring or getting withdrawn where the seller absolutely has to sell.

Life events. Job relocations. Divorces. Corporate REO. Estate liquidations. New-construction closings on the other end. Financial pressure. Sub-4% mortgages the seller is trying to preserve while relocating.

These sellers will list again. The only question is whether they list with the original agent, or with someone new. And in Q4 — as covered in earlier issues — most of your competitors are already checking out for the year. The prospecting window is open, and it's more open than it's been in a long time.

Use the "yet-to-be-sold listing" language. Systematic outbound contact. Phone first, mail second, door-knock third. And bring the AI CMA discipline described above to every appointment.

Five concrete moves:

One — build your AI pricing prompt today. Open your AI of choice. Ask it to write you the pricing prompt described above. Save the output. Test it on a listing you already priced. Refine until the numbers align. Ship the tested version to your next 10 seller leads.

Two — add the "which AI do you use?" question to your pre-qualification. Note the answer. Use that same AI to simulate the seller's likely fact-check before the appointment.

Three — pull current new-construction data for your market. Every active builder. Every current incentive. Current start-rate offers. Have this ready as competitive intelligence for every listing appointment where new construction is within 20 miles.

Four — reframe pricing conversations around monthly payment, not sale price. When a seller pushes for a higher list price, walk them through the resulting monthly payment at 6.75% and compare to the payment the buyer could get on nearby new construction at 4.75% for the same money. The math is often the whole argument.

Five — chase yet-to-be-sold listings this week. Systematic outbound contact. This is the largest available seller-lead source in the current market, and Q4 is the highest-yield prospecting window of the year for it.

Half of your sellers are already fact-checking your price with AI. That number is going to 100% within months, if it isn't there in your market already.

You cannot win by arguing with AI. You cannot win by trying to convince a fact-checking seller that ChatGPT is wrong. The AI is faster, more articulate, more confident, and available at midnight. In a fight, the seller sides with the AI.

You win by getting there first. Give the seller the professional prompt before they invent a leading one. Simulate their fact-check before the appointment so nothing surprises you. Seed your own AI with the market intelligence the seller's AI couldn't have. Show up with a CMA that makes theirs look like homework.

Then — and this is the deeper play — reframe the entire pricing conversation around monthly payment, not sale price, because that's what buyers actually shop on and what's actually determining whether your resale listing sells or your neighbor's builder listing does.

The agents who master this in the next 90 days are going to end 2026 with a full listing pipeline and a foundation for their best year ever in 2027. The agents who keep arguing with AI are going to lose those listings to the ones who don't.

Get to work.

Ready to stop guessing and start producing?

🎯 Start Premier Coaching (free trial): premiercoaching.com
💼 Build wealth with Tim's eXp team: whylibertas.com/harris
📲 Elite Coaching — text Tim directly: 512-758-0206

If you built the professional AI pricing prompt, deployed it to every seller before every listing appointment, and reframed every pricing conversation around monthly payment vs. new-construction competition — how many of the AI-fact-checked listings currently going to your competitors would be yours instead by January 1?

— Tim & Julie Harris

📬 Thanks for reading Harris Real Estate Daily. Share this with a colleague who needs clarity about where the industry is headed.

Forwarded by a friend? Sign up with just one click here.

Share this issue.