Big Idea

The argument in one line.

A wholesaler can hand an AI browser agent her exact search filters, comp rules, and offer math, and it will source, underwrite, and draft outreach for her, leaving only a final human approval step.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

You run a real estate wholesaling or land-flipping business and spend hours each week scrolling listings by hand.
You already have a tested set of filters and offer math and want software to run the repetitive parts instead of a VA.
You're evaluating AI browser agents for any lead-sourcing workflow, not just real estate.

SKIP IF…

You don't yet have a proven filter or offer formula, since the agent only repeats a process you've already validated.
You want a free solution end to end, since this setup pairs a paid AI agent with a separate paid underwriting tool.

TL;DR

The full version, fast.

Olivia Schremmer, a land wholesaler, walks the AI browser agent Muse through her entire Zillow pipeline. She teaches it her exact filters (4,000 sq ft to half an acre, $80,000 minimum, residential or duplex zoning, road access) and has it search four rotating markets on a fixed weekly schedule. Muse pulls candidate lots, hands them to the AI underwriting tool Assigns for comps, and calculates a maximum offer using only 'Perfect' or 'Great' comps, opening 10,000 dollars below that number. It then drafts personalized text offers through Google Voice at least 40,000 dollars below list price, logs every contact to prevent duplicates, and runs a Friday follow-up sequence plus a monthly follow-up blitz. Nothing sends without her approval.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →

Chapters

Where the time goes.

00:00 – 00:51

01 · Cold open: stop doing Zillow by hand

Olivia states the cost of manual sourcing, establishes her credibility (over $500K in assignments), and frames the video: teaching Muse her entire land-focused Zillow system.

00:51 – 02:34

02 · Step 1: teaching Muse the filters

She creates a fresh Muse account, names the assistant, connects it to Zillow, clicks through the human-verification prompt herself, and types in her first batch of filters: lot size, neighborhood quality, and price.

02:34 – 04:20

03 · Price rule and the multi-market schedule

She explains why $80,000 is her price floor, adds zoning/road-access/days-listed filters, assigns a different target market to each weekday (Dallas, Houston, Orlando, Atlanta), and Muse turns it into a live scheduled task.

04:20 – 06:18

04 · Step 2: connecting the AI underwriter

She switches from a flat 60% offer to a comp-based approach, connects Muse to Assigns.com, and has it pull a batch of candidate lots before it learns to run the AI comp calculator.

06:18 – 07:41

05 · Running comps and the MAO math

Muse is instructed to use only 'Perfect' and 'Great' comps and discard the rest, offer $10,000 below the calculated MAO, and the run returns a short list of survivors and tossed lots, further trimmed by a new $40,000-below-list filter.

07:41 – 09:54

06 · Step 3: texting realtors through Google Voice

Muse connects to Google Voice, pulls listing agents' phone numbers from Assigns, and drafts a short, deliberately vague cash-offer text for each one, following Olivia's on-market volume-strategy reasoning.

09:54 – 11:30

07 · The approval gate and duplicate tracking

Olivia has Muse log every agent/property contact to avoid duplicate messages, then reviews a batch of finished draft texts. Nothing sends until she explicitly approves it.

11:30 – 13:32

08 · The follow-up engine: sequence and blitz

She defines two follow-up cadences: a weekly Friday sequence to leads that went cold, and a monthly blitz that re-contacts every unanswered or rejected lead from the past 30 days, then recaps the full end-to-end system.

13:32 – 14:30

09 · Wrap-up: works for houses too

Olivia notes the same system works on houses with different filters, plugs her Assigns affiliate link, and the video closes on a chalkboard graphic summarizing the 3-step Zillow + Muse system.

Atomic Insights

Lines worth screenshotting.

A land wholesaler taught an AI browser agent her exact Zillow filters, comp rules, and offer templates, then let it run the entire search-to-offer pipeline.
The agent searches a different market each weekday, Dallas, Houston, Orlando, then Atlanta, so one person's pipeline covers four cities without manual switching.
The minimum price filter was set at $80,000 because anything below that rarely produces a meaningful assignment fee.
Only 'Perfect' and 'Great' graded comps are allowed into the offer calculation; anything graded 'Good' or lower gets the property thrown out entirely.
The opening offer is set $10,000 below the calculated maximum allowable offer, leaving room to negotiate upward in small increments without signaling a lowball from the start.
A separate filter requiring the offer to land at least $40,000 below list price exists specifically to leave room for both an assignment fee and an end buyer's margin.
Outreach texts are kept deliberately short and vague on numbers, because an over-explained message reads as amateur and invites more scrutiny, like proof-of-funds requests.
On-market wholesaling is treated as a volume strategy: send as many qualified offers as possible and let response rate do the filtering, rather than chasing each lead individually.
Every offer and every agent contacted gets logged, so the same agent or property never receives a duplicate message across separate automated runs.
The follow-up system runs on two cadences: a weekly Friday follow-up to leads that went cold that week, and a monthly blitz that re-contacts every unanswered or rejected lead from the prior 30 days.
No outbound message sends automatically. Every draft requires manual approval before it goes out, keeping a human in the loop on pricing and tone.
The same filter-and-offer logic applies to houses, not just land. Only the search filters need to change.

Takeaway

Automate the repeatable 80%, not the judgment calls

THE AUTOMATION PLAYBOOK

An AI browser agent can run an entire lead-sourcing pipeline end to end, filtering, underwriting, drafting outreach, and following up, as long as a human has already proven the filters and offer math work by hand first.

02Step 1: teaching Muse the filters

Give an automation your exact, already-proven filters rather than vague criteria, since the agent only reproduces a process you've tested, it doesn't invent a better one.

03Price rule and the multi-market schedule

Set a hard price or value floor before any other filter, since deals below a minimum threshold rarely justify the time spent working them.
Rotate a fixed schedule across multiple markets or segments so one pipeline covers more ground without you manually switching contexts each day.

04Step 2: connecting the AI underwriter

Move from a flat percentage-of-list offer to a comp-based calculation once volume justifies the extra underwriting step.
Connect a specialized tool for the part of the process that needs domain judgment (comps) instead of asking the general agent to estimate it.

05Running comps and the MAO math

Restrict any automated calculation to your highest-confidence inputs only; discard anything graded below your top two tiers rather than averaging in weaker data.
Open negotiations below your real ceiling number so there's room to move up in increments, instead of naming your best offer first.
Add a second, independent qualifying filter after the main calculation, so borderline results that technically pass still get cut if the margin is too thin.

06Step 3: texting realtors through Google Voice

Keep outbound messages short and light on specifics; over-explaining a number invites more scrutiny than it prevents.
Treat high-volume outreach as a numbers game and let response rate do the qualifying, rather than trying to hand-pick only the 'best' leads upfront.

07The approval gate and duplicate tracking

Log every contact attempt so a duplicate message never reaches the same person twice across separate automated runs.
Keep a manual approval step before anything goes out publicly, even when the rest of the pipeline is fully automated.

08The follow-up engine: sequence and blitz

Build two follow-up cadences, a frequent one for recent non-responses and a longer-interval one that sweeps up everything still unanswered, instead of relying on a single follow-up pass.

09Wrap-up: works for houses too

Design the system so the underlying logic transfers to an adjacent use case (here, houses instead of land) by changing inputs, not rebuilding the process.

Glossary

Terms worth knowing.

MAO (Maximum Allowable Offer): The highest price a wholesaler can offer on a property and still leave room for their assignment fee and the end buyer's profit margin, calculated from comparable sales.
Comp (comparable sale): A recently sold property similar in size, condition, and location, used to estimate what a subject property is realistically worth.
AI Comper: Software that grades comparable sales by quality, such as Perfect, Great, or Good, and uses only the strongest ones to calculate an offer price.
Assignment fee: The profit a wholesaler earns by signing a contract to buy a property, then assigning that contract to a cash buyer for a markup before closing.
Infill lot: A vacant parcel of land located within an already-developed neighborhood, as opposed to raw land on the outskirts of town.
Browser agent: An AI tool that can open a live web browser, navigate real websites, fill in forms, and take actions on a user's behalf, rather than just answering questions in a chat window.

Resources

Things they pointed at.

00:51toolMuse

00:00toolZillow

04:49toolAssigns.com ↗

08:25toolGoogle Voice

Quotables

Lines you could clip.

00:00

“If y'all are still doing every single part of your wholesale business manually going into 2027, you're wasting entirely too much time.”

cold-open thesis, states the stakes immediately→ TikTok hook↗ Tweet quote

03:40

“What repetitive thing am I doing in my business and ask, does it really require me? Scrolling on Zillow for three hours does not require me.”

reusable self-audit question, applies outside real estate→ IG reel cold open↗ Tweet quote

09:25

“I don't like to send super long texts. I don't make it overly complicated, because if you make it super long and explain every single number, it's going to sound very beginner.”

counterintuitive sales-copy advice→ newsletter pull-quote↗ Tweet quote

09:54

“Nothing sends until I approve it.”

tight, standalone line about keeping a human in the loop→ TikTok hook↗ Tweet quote

The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

metaphor

If y 'all are still doing every single part of your wholesale business manually going into 2027, you're wasting entirely too much time. A lot of you have seen my Zillow videos. You know I love low -balling realtors and sending out contracts.

But honestly, my time is becoming more and more limited. So when Muse came out, I decided to teach it literally my whole entire Zillow system. From my filters, to my offers, underwriting, it does it all.

And it even does my follow -ups. If you're new here, I'm Olivia Schramer, aka The Wholesale Girl. I've done over a half a million dollars in assignments.

and today I'm going to be showing you the exact tutorial on how I taught Muse to basically run my whole entire Zillow system for me. And we're specifically going to be focusing on land today. Just know that this does work on houses too.

So let's jump in. So step one is actually going to be teaching Muse what specifically I want to find. And this is where Muse probably saves the most time because you're not constantly scrolling through properties.

So before we do that, I want to show you guys my exact workflow that I have right now. So this is the current workflow that I have built out. i'll explain every single part of this but just know i have it doing quite a bit so i've created a brand new muse account so you can see exactly how easy this is and you can see that i'm starting from scratch so the first thing i'm going to tell muse to do is it needs to connect to zillow because obviously that's where it's going to find the properties next to zillow and it will have you verify that you are a human you're going to press take control of the browser obviously do this part So now I'm going to sign into Zillow here.

So now that we've gotten everything verified, now we're actually going to give Muse the filters that we want. I'm going to put all these in here and then we'll break them down. So I'm going to tell it, find me the following in fill lots slash land for the size.

I personally do 4 ,000 square foot to 0 .5 acres, half an acre. Then I want good neighborhood house on the left, the right, not on roads, not in the middle of nowhere and not across from commercial building or parking lot. Next one, personally price at 80 ,000.

minimum residential or duplex zone, cave road access, and listed for a minimum of 60. So if you need a more in -depth explanation of why specifically I chose those filters, be sure to go watch this video. I've talked about it before, but overall, I wanna target lots that are gonna make good wholesale deals.

So I obviously want a house on the left and the right. I don't want one in the middle of nowhere. I don't want 40 acres.

I don't want something just listed because chances are the seller isn't going to take a lowball offer. They're gonna weigh out their options and see how close they. get to their asking price and then i think one of the biggest things here is the fact that my minimum price point is 80 000.

i do that because if anything below that you start really not flirting with good assignment fees and it's kind of just a waste of time in my opinion so now that we have our filters we have to tell muse where we want it to look so personally i have it doing a different market each day it's making 20 offers i personally have mine targeting dallas houston orlando and atlanta however you can have a target whatever place you want now what i'm going to tell it is we are creating a workflow based on the filter given target bots slash land in the following market on the following So if you just want to stay in one market, that's fine.

I just like to dabble in different places because I can and I have other sources of leads coming in. So on Mondays, Dallas, Texas, Tuesdays, Houston, Texas. So don't do anything with that.

Don't screw up like I did. Wednesdays, Orlando, Florida. Thursdays, Atlanta, Georgia.

So now you may be thinking, Olivia, what do you do on Friday? We're going to get to that. Just wait.

So now I'm going to press go and it is going to find me those things. So instead of scrolling through 200 listings, it may only find me 10 that I actually care about. And that's more important to me than a bunch of random lots that I can't do anything with.

So it's important to sit there and think, what repetitive thing am I doing in my business and ask, does it really require me? Scrolling on Zillow for three hours does not require me. So.

systemize it so now you can see that it created my workflow so monday through thursday starting at 8 10 a .m it's going to look for the specific filters i wanted in the specific markets on the specific day and then it creates a scheduled task which is exactly what i wanted it to do now let's get into the number side of things so originally what i was having it do before i changed it i was having it offer 60 of the list price and i sat here and thought There's definitely a better way to do this.

And so if you have assigns, you can automatically hook it up to assigns and have it work through the AI Comper and offer based on that. So if you don't have assigns or a software that can get you to an ARV or new build comps or other land comps, that's okay. You can definitely have it offer 60 % or teach it to run numbers based off a formula.

However, I highly recommend assigns and I'm going to show you why. So before I have it actually go run numbers, what I'm gonna have it do is just find me 10 plots. that fit my criteria in orlando so this is for example purposes obviously it might take a little bit longer based on what specifically it's looking for but to speed up this process a little bit i'm just going to have it pull 10.

now while it's searching for those lots i'm going to have it connect to assigns .com i'm going to open this browser and if you have a signs you're going to take control of the browser and sign in So now once it's connected to assigns, while it's still searching for those properties, what I'm going to tell it to do is I log in.

I need to get familiar with the software. Once you find the lots, are going to the AI calculator to run. I only want the perfect and.

great if property if the lot doesn't have those throw it away if it does continue through the process get the offer the initial offer below or will need 10k below the mao So if you didn't understand what I just said, I'm telling it to get familiar with the signs and get familiar with the AI Comper because then I don't have to comp, but I'm also not just throwing random numbers out there.

So it's going to work through the assigned system to get to the offer. However, I only want it to use the perfect comps and the great comps, not the good comps and definitely not the bad comps. And then the initial offer is going to be 10 grand below the MAO because you don't want these realtors to know you're low -balling.

So you're gonna be able to go up in little increments. You know, you might be able to go up two. grand three grand all the way up until your mao and one thing you can actually do is watch muse go through the whole system and see what it's doing right now it's getting familiar with prop packet it's just getting familiar with the system it takes a little bit but it's worth it trust me i'll be honest i got a little impatient so it pulled some dallas ones and it only pulled seven this is what we're left with it's okay So there's one last thing that I'm going to add to this because it's going to filter it down even more.

So now what I'm going to tell it to do, I'm going to tell it my offer needs to be a minimum of 40 ,000 below list price.

So that's gonna eliminate some more. Here's the thing. Again, I got impatient and it didn't run every single property for me, but it will.

Now, the reason I specifically ask it to do 40 grand below list price is because I gotta leave room for an investor and myself for an assignment fee. So now here is what we are left with. Now that we have some properties, they fit our criteria, we need to move on to step three, which is making these offers along with the follow -up, which I'll explain the follow -up in a little bit.

We gotta send offers first. So what I'm gonna tell Muse to do is connect to... google voice okay in this part you will have to do manually so you can log in so once you've connected to google voice now we're going to tell muse how to get the phone numbers so we're going to tell it pull the phone numbers of the real from a sign.

If you don't have a signs, you can have it use the phone number on the listing. Now, while a signs is pulling those phone numbers, let's talk about what I actually send to the realtor. So while it's working on that are going and this.

Hey, that's the template I'm going to use. So all this says, hi, this would be the agent's name. My name is Olivia.

I'm interested in your listing on the street. I reviewed the numbers and spoke to my lender. I'll be up.

offer cash with 21 day close this is something your client is interested in let me know and i can get a contract sent over i don't like to send super long books i don't make it overly complicated because if you make it super long and you explain every single number it's going to sound very beginner and i don't want to do that because then they start asking for proof of funds and asking just a lot more questions in general so might as well avoid that altogether i've said this before but in my opinion on market is a volume strategy so i just want to send out as many offers as possible and see what kind of responses i get and then that's when i'll send contracts or move on to the next one or follow up which again we're going to get into here shortly as you can see it saves it as a template and then once it's done finding phone numbers it is going to draft out the whole text message and i think the best part about this is nothing sends until i approve it if you're still unsure about numbers or there's just something that doesn't seem right you can go back and rework it and one more thing i do tell it to do is keep track

of every agent slash property message. I don't want to send duplicate messages over and over, plus we are going to create a follow -up strategy.

So while this is working on finding agents' phone numbers and drafting my text messages, let's talk about what Friday's for. Because if you remember, what I had it do is send offers Monday through Thursday, and then I left Friday out. Friday is specifically my follow -up days, and so what I'm having it do next is it's going to follow up with the leads that didn't respond from the week prior, or the ones that just flat out said no.

And I'll tell you what exactly I say. So now I got all five of my phone numbers. I'm gonna tell it to draft the messages.

And now here are all the drafts that I have for each property. And it already has the agent's phone number. Now I'm not actually going to have it send these text messages in this video just because I've already texted some of these properties before.

We're not going to mess with that. But all you would have to do is say send messages. And then it'll go into your Google voice and send those messages from there.

So now all of our messages have been sent out. We need to talk about the followup sequence because once you send offers, it's not just done. I want to know at the end of the week, which leads didn't answer, which ones responded and which ones just said no or something similar to no.

So now we're going to create the followup sequence and the followup blitz. So the difference between the two is the sequence is going to tell me about the week prior. It's going to tell me which agents said no and which agents just flat out didn't respond.

And then it's. to send out follow -up messages and the blitz at the end of the month what i'm having it do is take the whole month of leads i contacted and again whichever agent said no or didn't respond we're going to send follow -up messages to them as well so for the follow -up sequence track of which agents didn't respond said no or too low from the prior send them follow up make sure the proper properties are still available then i'm going to tell it send quotation hey still very property on street would your client reconsider my offer so that's the follow -up sequence now let's tell it about the blitz for the blitz go back all the messages that whatever agent didn't respond said no or too low we are going to send them one more follow -up this happen on the last of every month so that's going to keep my follow -up clean and again no messages go out until i confirm it so here is an overview of what i just talked about so it pulls the agent's phone number from assigns it sends them a text message after it gets all my offers in place and once i confirm it and then if it's the same listing agent for several properties it's going to combine it into one text message rather than sending the same thing over and over and over again just for different properties that's going to keep it really clean but also it's going to keep track of every single offer that was sent and it's going to create a follow -up sequence for the previous week and it's going to message those on friday so friday

are follow -up days and then the last week of every single month it's going to take every single offer and compile the agents and send out follow -ups again so that final week is pretty much follow -up week so i really want to preface that having a software that pretty much does all the work for you makes it super helpful along with muse so if you have a signs and you have muse it's just a deadly combo.

You don't have to focus on underwriting. You don't have to train Muse to come up with offers, which means things are going to get done faster because Assigns has the AI Comper. If you're interested in Assigns, click that link below and you'll get 50 % more AI searches.

But overall, you can do this for free. You don't necessarily need a software. It just makes life a lot easier.

It can pull phone numbers from places like hard .com and you can teach it to comp. It's just going to take a little bit more time. And also this does work with houses.

You just need to change the filters. I have a sequence going. I have a whole workflow.

going for houses too. I just prefer land. So that's pretty much it.

All I have Muse do is find the property based on my filters. Then I have it come. Then I have it send my offers.

And then it has its own follow -up sequence. So it's all hands off. Other than the fact that I have to approve the text messages before they're sent.

So going into 2027, eliminate some of those things that don't necessarily require your time. And Muse is a great way to help you with that. If this video was helpful, as always, be sure to like, comment, and subscribe.

And we will see you in the next one.

The Hook

The bait, then the rug-pull.

Olivia Schremmer, who goes by The Wholesale Girl, opens by naming the cost of doing this by hand: hours of scrolling Zillow for a business she says has already produced over half a million dollars in assignments. The rest of the video is her handing that exact process, filters, comps, offers, and follow-up, to an AI agent named Muse.

Frameworks

Named ideas worth stealing.

14:13model

The 3-Step Zillow + Muse System

Find the deals
Run the numbers
Make the offer + follow up

The organizing structure of the whole video: filter-driven sourcing, comp-based underwriting, then templated outreach with a built-in follow-up cadence, all handed to an AI browser agent.

Steal forany repetitive lead-sourcing workflow (real estate or not) where the filters and math are already proven by hand

06:18concept

MAO offer rule

Use only Perfect and Great comps
Open $10,000 below MAO
Raise in small increments toward MAO

A fixed formula for turning a graded comp set into an opening cash offer that leaves negotiating room without revealing the lowball upfront.

Steal forany offer-negotiation workflow that wants a defensible, repeatable opening number

07:25concept

$40K-below-list filter

Offer must be at least $40,000 below list price

An extra post-comp filter that guarantees enough spread for both the wholesaler's assignment fee and the end buyer's margin before a lead is even worth texting.

Steal forqualifying leads before spending outreach time on them

CTA Breakdown