What they are looking for
Buy or rent intent (10), target area (15), budget (20), and property type (10).
A self-initiated local prototype that checks buyer and renter enquiries, explains which details are present, and gives an agent a synthetic listing match to review.
A buyer or renter may describe a budget, area, property type, timing, or financing in an unstructured message. An agent then has to find the missing details, decide what to clarify, and compare the request with available listings. This prototype brings those steps into a reviewable local workflow.
The names, enquiries, and property records used in the walkthrough are fictional. The examples below describe tested workflow behavior.
Points represent whether useful enquiry information was provided. They do not predict conversion or determine access to housing, a viewing, finance, or a property.
Buy or rent intent (10), target area (15), budget (20), and property type (10).
Timeframe contributes 5–20 points; financing information contributes 8–15 points.
A usable contact method and a provided name each contribute 5 points.
The prototype labels 70+ as hot, 40–69 as warm, and below 40 as cold. These are internal completeness bands for demonstration only.
The active local n8n workflow was checked with synthetic enquiries. The workflow editor contains 25 nodes; validation checked 23 connections and 27 expressions with zero errors and one expected warning for the disabled Telegram node.
HTTP 200, readiness score 100 (hot), one fictional match (ROM-S-001), and a simulated notification.
HTTP 200, readiness score 45 (warm), one fictional match (ROM-R-001), and missing information returned for agent follow-up.
The local llama3.2:3b model extracted the requested fields; deterministic scoring returned 100 and matched ROM-S-001.
Missing contact returned HTTP 422. Resubmitting a stable lead ID updated one matching Data Table row instead of creating a duplicate.
With an unavailable model tag during setup, the request still returned HTTP 200, used mock_fallback, and included agent-review guidance.
Execution 547 succeeded and included lead fields, score reasons, a listing match, and a suggested next action. No email was sent in this preview.
This screenshot shows the connected workflow structure, including local Ollama, mock notification, optional Gmail, and the disabled Telegram node. It is a workflow-editor capture, not an execution result or proof of message delivery.


This static map is drawn from the verified workflow structure for easier reading. It is not another screenshot or an execution log.
This is self-initiated work with fictional listings and synthetic enquiries—not client work, production software, or evidence of business results.
Matches come from fictional local records. No client CRM, listing feed, property portal, or live availability system is connected.
A readiness band is not a measure of a person’s worth or eligibility. An agent must check the request, listing, availability, and appropriate follow-up.
n8n, local Ollama, and local Data Tables were exercised. A synthetic Gmail send was verified; the edited HTML was previewed with mock mode. Telegram was not tested.
No public deployment, real estate CRM or inventory integration, consent and retention design, access controls, monitoring, repeat Gmail delivery, or production reliability was verified.
DM “CALLS” on Instagram with your business type and how you currently handle new enquiries.