How I built one system where every lead, call, email and listing for a busy brokerage lives on the right record.
By Jude Horak | Updated November 15, 2026
It looks like the team's own day. For this brokerage that means each boat inquiry gets its own record on one contact card, routed to the listing broker. AI replies and follows up, then hands off when a person is needed. Calls and emails log themselves, and the same listing data feeds the marketplaces.
listing records a day sent from the CRM to the major boat marketplaces
per person, with a separate record for every boat they ask about
on the company domain is read, and matching emails land on the right record
before any AI message a broker triggers, so they can read and edit it
A marketplace inquiry, a website form or a call to the phone line.
Joined to the person's card and assigned to the broker who lists that boat.
Answers within minutes, asks what times work, and flags the broker when needed.
Calls, texts and emails from the record, all logged onto the deal.
The same listing data feeds the marketplaces and comes off when the boat sells.
The brokerage sells boats under several brands, and leads come from everywhere: the major boat marketplaces, its own brand websites and a phone line. One buyer often asks about more than one boat, and each boat can belong to a different broker. Before this build the team worked out of a help-desk tool and a web of automations that passed leads around.
So I modeled the business as it is. A person has one contact card. Each boat they ask about is its own record, assigned to the broker who lists that boat, with its own email thread. Repeat inquiries are matched by email and phone so the same person does not appear twice.
A dashboard of what needs them today: overdue and upcoming tasks, plus cards built from what is happening on their records. Red means a person has to act. Yellow means the AI raised something. Green means a showing is ready to confirm.
Buyers and sellers each have a board and a list. The seller board runs from a new seller through working, opportunity and a signed listing agreement, then on through the deal stages to closing. Booked showings land on the broker's own calendar.
There is also an in-app assistant that can answer questions about their tasks, buyers and appointments. It only sees what that broker is allowed to see.
When a buyer inquiry comes in, the AI writes back within minutes from the assigned broker's own mailbox, so replies thread like normal email. It answers questions about the boat, and if the buyer has a phone number it can text too. When the buyer wants to see the boat, it asks what times work. Once they name real days and times, the broker gets a card with those times to tap and confirm.
On the seller side, owners who have not answered get short follow-up notes on a set schedule. The AI reads the texts, emails, notes and calls on a seller's record. When it is confident the owner wants to list, it moves the card forward.
Where a person approves: whenever a broker triggers an AI message, they see a preview they can edit before it goes. Each buyer record can be set to draft only or auto-send. When a person takes over a conversation, the AI steps back.
Calls to the phone line create or update the caller's record, and the recording and transcript can be played and read right on it. Brokers call and text buyers and contacts straight from a deal through the CRM's phone, and those calls and messages show up in the deal's activity.
Every mailbox on the company domain is read automatically. If an email is to or from a known lead or contact, it is logged onto their record. Email from anyone else is left alone. Website contact forms land on the person's record with a task for its owner.
Listings live in one place, and the CRM builds a feed for each of the major boat marketplaces in the format each one asks for. That comes to about 2,000 listing records a day. Broker contact details are stripped from descriptions where a marketplace forbids them, listings are routed to the right regional account, and sold boats come off. A manager can hold any listing back from the feeds.
I do. I am Jude Horak, the engineer behind Everyday Software in New Lenox, Illinois. For this brokerage I built the CRM, connected it to their phone line, mailboxes, website forms and marketplace feeds, and I still run and improve it.
When the work does not match the default objects. Here one person can ask about three boats held by three brokers, a seller can also be a buyer, and every listing has to reach several marketplaces with different rules. A generic CRM would need workarounds for each of those. A custom one models them directly.
Yes, with clear limits. In this build the AI answers buyers and follows up on its own, but it hands the conversation to the broker for anything only a person can do, it stays quiet when a buyer is upset, and a broker sees a preview of any AI message they trigger before it goes.
The brokerage ran on a help-desk tool and a web of automations. I imported the open deals, notes and documents, kept the past call recordings, and pointed the buyer inquiry automations at the CRM so new leads land there and nowhere else.
It depends on scope. The things that move it most are how many lead sources need connecting, whether AI replies are part of it, how much old data has to be moved, and how many outside systems, like marketplaces or phone lines, it has to talk to.
Tell me how your sales day actually runs and I will show you what a CRM built for it would do.
We use optional analytics and marketing cookies to understand visits and enquiries. Your demo works with either choice. Privacy policy