Manufactured housing AI case study

How a 2,500-deal operator built an AI edge competitors cannot copy.

Dan and Dan already knew how to buy mobile homes better than almost anyone. The unlock was turning that hard-won judgment into a 24/7 intake and offer system that helped nearly triple net operating income without adding headcount.

3x NOI growth year over year
2x+ Sales volume with the same team
2,500+ Real transactions behind the system
YouTube video 1
Founder interview Paste YouTube ID after upload
YouTube video 2
System walkthrough Paste YouTube ID after upload

They did not need generic AI. They needed their own operating judgment in software.

Dan Leighton and Dan Paton have bought and sold manufactured homes for over two decades. They know the market cold: the right parks, the right prices, the right questions to ask a seller, and the early red flags that make a deal a waste of time.

24/7 Lead intake and seller qualification
0 Additional headcount required for first touch
1 Proprietary offer engine trained on their process

"Speed to lead has definitely made a difference. Once in a while we get an FU. That's it. The rest of the time, it's working."

Dan Paton, co-founder, EZ Homes

Casey handles the first mile of every deal.

Working with a custom AI development team, EZ Homes trained an AI intake agent on their exact acquisition process: how to qualify leads, assess home condition, identify park ownership, and surface a calibrated offer range before a founder picks up the phone.

What Casey does today

  • Captures inbound seller leads instantly and begins qualification without human first contact.
  • Collects condition, address, park identity, upgrades, flaws, and seller motivation through natural conversation.
  • Auto-calculates a high and low offer range from a database built on 2,500+ closed transactions.
  • Flags high-value parks, known ownership groups, and deal killers before founder time is spent.
  • Turns archived Facebook Marketplace conversations into a reusable buyer and seller pipeline.

The moat is not the chatbot. The moat is the transaction history behind it.

Most companies can copy a prompt. They cannot copy thousands of real transactions, years of park-level judgment, and a battle-tested offer process. That is what makes the system compound: every qualified lead strengthens the operating data that guides the next one.

"We are a bit of a unicorn out there. A couple of jabronis who figured out how to do something that nobody else is doing, and now we're building the system that locks in that edge forever."

Dan Leighton, co-founder, EZ Homes

How this kind of AI project actually gets built.

Step 01

Map the founder judgment

Extract the questions, red flags, scoring rules, and offer logic the best operator already uses instinctively.

Step 02

Build the workflow

Connect intake, CRM, data enrichment, qualification, offer calculation, and handoff into one reliable operating flow.

Step 03

Improve with real usage

Review misses, edge cases, and conversion data so the system gets sharper instead of becoming shelfware.

If your best people are still doing repeatable judgment work manually, there is probably an AI system hiding in plain sight.

Bring the workflow you want to scale: lead intake, qualification, quoting, follow-up, reporting, or internal decision support. We will help identify the fastest path to a useful system, not a science project.

Request a walkthrough