Case study · Commercial real estate · Automotive
Calling 36,000 shop owners without losing one
A commercial brokerage that sells, buys and leases automotive properties and businesses prospects owners by phone. I built its GoHighLevel account around one loop: call, pick a disposition, and let the CRM handle the next step. An AI agent verifies the list first, a custom object links owners to all their properties, and seller questionnaires produce a valuation before the broker calls back.
The system
Call, disposition, next step
Owner lists are cleaned and verified before a broker dials. Each call ends with one disposition, and that disposition decides where the contact goes: back to the dialer, into the deal pipeline, to inside sales, or off the list. Interested owners get a questionnaire, and the answers produce a valuation and listing-agreement data.
Comes in
GoHighLevel
Goes out
System 1 · Outbound
Live-caller power dialer
Thousands of owners to call, bad rows in every list, and no rule for who calls whom.
- Three triggers load the dialer: a calling stage in the outside-sales pipeline, a calling stage in the inside-sales pipeline, or the "unverified" tag coming off after an AI call.
- An "appropriate for calling?" filter drops broker-list contacts, corporate chains, trucking companies and marinas before anyone dials.
- A contact with an assigned caller goes back to that caller. Everyone else gets a 50/50 round-robin split, a follower, a dialer-queue tag and a saved assignment.
- Regional caller numbers give owners a local area code.
System 2 · Routing
Ten dispositions, ten next steps
After each call, callers had to remember to update the pipeline, write notes and requeue the contact.
- Each call disposition is its own trigger: Interested, Financials Requested, No Answer, Voicemail, Not Interested, Needs Research, Not Available, Corporate Owned, Answered – Not Owner, Do Not Call.
- Interested and Financials Requested remove the unverified tag, write a note, then find or create the opportunity in the right stage and add followers.
- No answer, voicemail, not interested go back to the live list or to a manual caller.
- Do Not Call, corporate owned and not-owner come out of the dialer.
System 3 · AI
An AI agent verifies the list first
Purchased lists had wrong business names and owners who had sold years ago.
- A Voice AI agent makes short verification calls to unverified contacts before a broker spends time on them.
- The prompt is strict: confirm the business name, never trust the stored contact name, ask who handles property decisions, and only mark Interested after a clear yes.
- In-call actions trigger the disposition workflows and save the agent's outcome to a contact field.
- A knowledge base answers "who are you?" questions. Verified owners land in New Lead From AI Caller.
System 4 · Data model
One owner, many properties
A contact record holds one address, but shop owners often own several locations.
- A Properties custom object stores each location as its own record, keyed on a unique property address.
- A many-to-many association links properties and contacts: one owner can hold several locations, and one location can have an owner, an operator and a landlord.
- Brokers open one contact and see the whole portfolio, which is what sale-leaseback and multi-site conversations need.
System 5 · Qualification
Questionnaires move the deal
Brokers chased owners for financials and details before they could value anything.
- Separate questionnaires for business sale, business purchase, property sale, property purchase and leasing.
- Clicking the link in the email moves the opportunity to Financials / Info Requested straight away.
- Submitting sends a thank-you email and finds or creates the opportunity in Questionnaire Received.
- Buyer and tenant requirements are saved in their own field folders, so brokers can match buyers to listings.
System 6 · Valuation
A valuation before the callback
Owners wanted a number, and brokers had to work it out by hand first.
- When a Business Sale questionnaire comes in, the workflow reads net income and applies the multiple for its tier.
- The result is saved to Estimated Business Valuation, and the lead broker and contact owner are notified.
- Listing-agreement fields (entity, price, inventory, compensation, dates) live in their own folder and a second custom object, so the agreement is drafted from CRM data.
Structure
A 22-stage pipeline from research to escrow
The sales pipeline follows a brokerage deal: research the owner, reach them, get financials, value the business, sign a listing, find a buyer and close. A second, two-stage pipeline holds contacts the AI caller has verified for the inside-sales team.
Sales pipeline by stage
319 open opportunities · stages with deals
Busiest workflows
Total enrollments, October 2026
System 7 · Deliverability
Cold email that stays out of spam
Sending tens of thousands of cold emails from the main domain would damage its reputation.
- The intro email explains the brokerage's niche and links to the questionnaires. It goes out through GHL's outreach email action.
- Warm-up campaigns run on five secondary sending domains, so cold volume never touches the main domain.
- STOP replies set DND on SMS, email and calls automatically, and owners who call back get a text thanking them.
Organization
Folders a new admin can follow
Workflows are grouped by job: AI calling, form-link triggers, maintenance and field mapping, manual calling and dispositions (live callers, inside sales, outside sales), marketing, pipeline automations and website forms. Background jobs like contact-ID sync, voicemail alerts and lead validation sit in their own folder, away from the sales logic.
| Asset | What's in the account |
|---|---|
| Calendars | 13: a booking calendar for each broker, plus an A2P compliance calendar |
| Funnels | A2P compliance (7 steps), internal contact entry, seller document upload, valuation template |
| Forms | Business listing intake, seller documents, interested property, contact entry, contact us, careers |
| Email templates | Intro email, valuation emails, document sent, newsletter welcome, thank-you |
| AI agents | Outbound verification Voice AI, a Conversation AI assistant, two knowledge bases |
How it's built
Tools and design choices
Callers only pick a disposition
Notes, pipeline moves, requeues and removals all happen on their own.
Bad data never reaches a broker
Exclusion filters and AI verification calls clean the list first.
The data model fits the business
Owners and properties are linked many-to-many, so a whole portfolio shows on one record.
Brokers call back with a number
The questionnaire produces a valuation before the first serious conversation.