Fig. 7 Buyer pipeline
A bilingual buyer pipeline with AI property matching
52 GHL workflows that take buyers from first contact to closed deal in English or Spanish, plus a custom app that matches each buyer to the right homes.
- Client
- US real estate investor
- Industry
- Real estate investing
- Region
- United States
- Discipline
- GHL systems, Custom apps, AI workflows
Drawing
Trigger
New buyer lead
SMS + wait
Discovery texts
If / else
Qualified
Calendar
Showing booked
Document
Term summary sent
Done
Contract drafted
The problem
The business needed one system for buying and selling homes: buyer intake, property acquisitions and deal follow-through. Many leads speak Spanish, and matching each buyer to the right home was done by hand.
Built with
- GoHighLevel
- CloseBot
- OpenAI GPT
- React
- TypeScript
- Supabase
- Airtable
- Vercel
- Apify
- Mapbox
What I built
- 0152 published workflows in 5 folders: buyer lead intake, buyer home disposition, property acquisitions, conversation AI and supporting automations, 956 steps in all.
- 02Bilingual follow-up, with GPT steps translating messages for Spanish-speaking leads and a 10-day SMS drip in both languages.
- 03An AI assistant on CloseBot that qualifies buyers and captures income, liabilities and down payment, plus a custom-code check that flags affordability risk.
- 04A custom matching app that pairs buyers with properties, sends listings one at a time or in bulk, and records only the sends that actually delivered.
- 05Sold homes drop out of public listings and matching automatically, and creative-financing listings are pulled in from Zillow.
Build notes
Sold homes now leave listings and buyer matching automatically
One toggle marks a property closed. It gets a Sold badge, drops out of public listings and buyer matching, and keeps its deal history. Also switched the creative-financing Zillow search to use Zillow's own keyword filter.
Every screen now agrees on what was sent to whom
Sent status now comes from one shared rule, so counts match everywhere. Matches record only deals that happened, and send screens report partial failures honestly instead of a blanket success. About 600 lines of code came out.
Passing on a home no longer deletes the match history
Operators can mark a property not interested for a buyer, one at a time or in bulk, and restore it later. Also stopped GHL silently dropping buyer income and down-payment updates, and flagged buyers gone quiet for 30-plus days.
Next drawing
Monthly client reports drafted by AI from Asana