Video summary

I Built a Real Estate AI Agent with n8n & Sold It for ₹33,000

Main summary

Key takeaways

Business

Business goal / pain point

  • Real estate and interior businesses often receive many inbound leads but lack a follow-up system, so leads go to competitors.
  • Proposed solution: build an automation “lead follow-up workflow” that starts immediately when an inquiry form is submitted.

Offer & pricing / sales motion

  • Created a reusable n8n workflow:
    • Lead → scoring → messaging → calendar booking
  • Sold it to a real estate developer for ₹33,000, positioned as a beginner-friendly “minimum ₹30,000” offer.
  • Reusable template playbook
    • After initial setup, the template can be sold repeatedly by swapping client-specific details (same workflow, different data).
  • Marketplace/channel for selling digital assets:
    • Super Profile — sells PDFs/images/videos/prompts and “digital products” via a dashboard workflow.

Core workflow (end-to-end operations playbook)

1) Trigger (inbound capture)

  • When a customer submits an inquiry form on the client website:
    • A Webhook/API POST starts the workflow.

2) Data normalization

  • A JavaScript node cleans and formats incoming lead data into valid JSON.

3) CRM storage

  • Google Sheets is used as the CRM.
  • The leads table includes fields such as:
    • client name, number, email, budget, location, property type, lead score, recommendation(s), response.

4) Duplicate prevention

  • Google Sheets “Get Rows” checks whether a lead with the same contact number already exists.
  • If duplicate:
    • set isDuplicate = true
    • branch to a duplicate-handling message
  • If not duplicate:
    • continue to scoring and follow-up.

5) Messaging infrastructure

  • Twilio configured for WhatsApp
    • Sends WhatsApp messages to the lead.

6) AI lead scoring & qualification

  • Uses an LLM chain via Ollama model:
    • Qwen 3 Coder 480B
  • Prompt role: “expert real estate lead qualifying agent”
  • Scoring rules (presence-based):
    • Budget mentioned: +20
    • Location mentioned: +20
    • Property type specified: +20
  • Output requirements:
    • Returns structured JSON only (to support deterministic routing).
  • Example outcome:
    • A lead can receive a score of 100 when multiple intent signals are present, alongside:
      • a follow-up recommendation/message
      • qualification status.

7) Lead lifecycle routing (hot / warm / cold)

  • A Switch node routes based on lead score:

    • Hot
      • score ≥ 20 → booking flow
    • Warm
      • score between 50 and 80 (“Vam” in subtitles) → follow-up recommendations (no calendar booking)
    • Cold
      • score < 50 → 2 recommendations + “personalized report” style message

Note: Thresholds appear somewhat inconsistent in subtitles, but the intent is tiered messaging by score.

8) Booking for Hot leads

  • Google Calendar
    • Fetch available slots (Get Many Events), with Return All enabled.
    • Apply office-hours constraint via JavaScript:
      • 9 to 6 (subtitles mention 9 to 5 in code explanation)
      • code checks only slots within business hours
    • Create event on chosen slot and add:
      • lead as an Attendee
      • AI/lead details in Description
  • Confirm booking via WhatsApp using Twilio.

9) Follow-up for Warm / Cold leads

  • Warm (no booking):
    • WhatsApp message with recommendations for 3 properties
  • Cold (no booking):
    • WhatsApp message with recommendations for 2 properties
    • plus a line that the team will send a personalized property report

10) CRM updates

  • After messaging/booking:
    • workflow writes updates back to Google Sheets:
      • lead status, lead score, recommendation(s), response.

Frameworks / playbooks explicitly implied

Lead scoring framework (rule-based + LLM justification)

  • Score is driven by presence of buyer intent fields:
    • budget, location, property type (and implied completeness such as timeline)
  • Enforced JSON-only structured response for consistent workflow routing.

Lifecycle segmentation

  • Hot → calendar booking + confirmation
  • Warm/Cold → curated property recommendations

Automation template & repeatable productization

  • Build once as an n8n template → customize → sell repeatedly.

Key metrics / KPIs and numeric targets mentioned

Project economics / ROI claims

  • Workflow sold for: ₹33,000
  • Claimed owner cost per lead: ₹2 to 3 lakh (described as “approx.”)
  • Conversion scale assumption: 10–12 leads converted via automation
  • Profit uplift claim: ₹20–30 lakh in 2 months
  • Example “payment” claim:
    • even if client pays ₹1 lakh, “no loss” (implied ROI justification)

AI scoring example

  • Lead score example: 100 when multiple intent signals exist.

Operational window

  • Booking only during 9 to 6 (subtitles: 9 to 5 in code explanation).

Recommendation counts

  • Warm: 3 property recommendations
  • Cold: 2 property recommendations

Concrete technical components / implementation checklist

n8n nodes described

  • Webhook trigger (POST)
  • Code (JavaScript) to clean/format data into JSON
  • Google Sheets node to check duplicates via contact number
  • IF/Switch routing
    • handles duplication and lead score tiers
  • LLM chain (Ollama Qwen 3 Coder 480B)
    • produces lead score + message in JSON
  • Google Sheets CRM update
    • append/update CRM records (columns include client info + lead score + recommendation + response)
  • Google Calendar nodes
    • list events to find available slots
    • create event with description + attendee
  • Twilio
    • WhatsApp sending + QR / webhook URL configuration

Delivery mechanics

  • Creator offers: share JSON workflow if viewers comment/DM (to import quickly).

Lead generation tactic (high-level)

  • Uses a browser extension “Map Lead Scraper” to scrape real estate leads for a city.
  • Produces CSV/Excel output.
  • Example shows 43 leads including:
    • names, addresses, contacts, websites
  • Mentions ability to email leads automatically (details deferred to another video).

Concrete recommendations (business execution)

  • Productize the workflow as a digital template
    • sell to nearby real estate/interior businesses.
  • Position the ROI around:
    • speed-to-lead (instant WhatsApp confirmation)
    • no missed leads (duplicate checks + automated routing)
    • tiered follow-up (booking only for hot leads; scalable recommendations for warm/cold)
  • Use Google Sheets as a lightweight CRM to track:
    • lead status changes, recommendations sent, and booking outcomes.

Sources / presenters

  • Presenter/creator: the YouTube video narrator (not explicitly named in subtitles).
  • Tools/third-party sources mentioned:
    • n8n
    • Google Sheets
    • Google Calendar
    • Twilio (WhatsApp)
    • Ollama (Qwen 3 Coder 480B)
    • Map Lead Scraper extension
    • Super Profile (digital product marketplace)

Original video