Loading
Loading
A blueprint for a brokerage lead system: every buyer and seller inquiry gets an instant reply, an AI assistant qualifies by text, ready buyers book showings on the right agent's calendar, and everyone else enters long-horizon nurture, with agents approving each handoff.
Solution blueprint: a system Callimachus Forge can architect for your operation. Not a case study, and not a claim of past deployment.
At a glance
Stack candidates
n8n · GoHighLevel · Twilio · Calendly · Google Calendar · Gmail · Slack · Postgres
System run
A simulated run of the exact steps documented on this page, from first trigger to final handoff. Nothing here is live client activity.
Simulation · Loops
The problem
Speed decides who gets the client. A buyer who fills out a form on a Saturday night usually inquires on several homes at once and works with whoever answers first. But the people who should answer are in showings, in the car, or off the clock. So paid portal leads sit for hours, follow-up depends on whichever agent happens to be free, and leads who say they are thinking about next spring are forgotten entirely. The brokerage pays for lead flow at the top and loses most of it to slow response and inconsistent follow-up, and no one can say which source actually produces closings.
A residential brokerage or team runs on a steady mix of inbound: portal inquiries, website forms, sign calls, open house sign-ins, and referrals. Leads arrive at all hours and route informally, a round-robin text thread or whoever is on floor duty. Some teams staff an inside sales agent to call new leads, but coverage ends when that person goes home. Agents are independent operators with their own habits, so follow-up quality varies widely across the same brand. The CRM exists but is only as good as what gets typed into it, and long-timeline sellers, the most valuable future listings, are the least likely to be tracked.
The manual process
The automation opportunity
The blueprint puts a system on the front door. Every inquiry, from any source, gets an immediate personal text and email that references the actual property or request. A conversational assistant then asks the same qualifying questions a good inside sales agent would: budget, financing, timeline, must-haves. Ready buyers are offered real showing times from the right agent's calendar and booked on the spot. Sellers are captured and routed to a human quickly. Everyone else enters a patient, months-long nurture that watches for a reply. Agents step in at defined points, always with the full conversation in hand, and the brokerage gets clean records of what every lead source actually produced.
Architecture
Every step is typed: triggers, AI decision points, business logic, actions, data, and the human checkpoints that keep judgment where it belongs.
AI components
Human checkpoints
n8n webhooks receive website forms and ad leads. A Gmail parser reads portal notification emails and pulls out name, phone, property, and message text. Sign calls and open house sheets post to the same intake endpoint.
The intake flow searches GoHighLevel by phone and email. Existing contacts get the new inquiry appended to their timeline; new ones get a record with source, campaign, and property of interest stamped on it.
A language model tags each inquiry: buyer, seller, renter, or vendor spam. It extracts price range, neighborhoods, timeline, financing signals, and for sellers the property address, into structured CRM fields.
Within the same run, GoHighLevel sends a text and email from the assigned agent's number: it acknowledges the specific property or request and asks one natural qualifying question. Spam-tagged leads are held instead.
An AI assistant works the SMS thread: it gathers budget, preapproval, timeline, and must-haves, and answers listing questions from a vector database of current inventory. Uncertain questions are escalated, not guessed.
Rules score each lead from the extracted fields: ready to tour or list, warm but not ready, or unresponsive. Each score maps to a branch: booking, nurture, or a spaced retry sequence that stops after defined attempts.
For ready buyers, the flow reads the routed agent's Google Calendar via Calendly and offers real open slots by text. A booking writes to the calendar and triggers confirmation and reminder messages.
Ready leads, seller inquiries, and escalated threads land in a Slack queue with the transcript and extracted profile. An agent approves the handoff, takes over the thread, or corrects the classification.
Warm leads enroll in a GoHighLevel cadence matched to their stated timeline: area market updates and periodic AI-drafted check-ins. Any reply exits the cadence and re-enters live qualification.
Every message, score change, booking, and handoff writes to Postgres alongside the CRM, keyed by lead and source. This gives the team a clean history per contact and true source-to-closing attribution data.
The agent opens each showing with a short brief: who the lead is, what they want, what was said, and how they came in. A recurring digest shows lead volume, response coverage, and pipeline stage by source.
Engineering notes
Benefits are qualitative by design. We do not publish numbers we cannot verify.
The free audit maps your actual tools and process against this architecture, and tells you honestly what is worth building.