Independent real estate agents were managing their entire pipeline manually. New leads from web forms arrived in an inbox and had to be manually evaluated, scored by gut feel, copied into ClickUp, and followed up on — whenever the agent had time.
High-quality leads sat uncontacted for hours. Low-quality leads consumed the same effort as serious buyers. When a deal closed, the agent spent the first day of the new engagement creating Google Drive folders, setting up ClickUp task lists, drafting a client welcome email, and sending an internal brief to the team — four separate manual tasks that added up to 60–90 minutes of administrative work before any real work began.
The pipeline itself was the bottleneck, not the deal flow.
Two n8n workflows that together cover the full lead lifecycle. Workflow 1 handles lead intake and AI qualification — scoring, routing, and follow-up. Workflow 2 fires on deal close and provisions the entire client project in under 60 seconds.
Every inbound enquiry hits a webhook. A single Claude Opus 4.5 agent normalises the raw form data, scores the lead 0–100 against a qualification rubric (budget fit, timeline, property type, motivation signals), and generates all downstream content — task name, email body, Slack message, and calendar details — in one structured JSON response. A Parse node then enforces the routing tier deterministically from the numeric score, overriding any free-form AI classification.
Leads scoring 70 or above are created as high-priority Active Leads in ClickUp with an AI-written qualification comment. A branded HTML welcome email fires immediately, a 30-minute discovery call is booked in Google Calendar for the next business day, and the agent gets a Slack alert with the full lead summary. Leads below 70 go to the Nurture list with a value-first follow-up email and a 14-day due date.
When a deal status changes to 'Won' in ClickUp, a webhook fires Workflow 2. The full ClickUp task JSON is fetched — including client name, company, email, budget, and requirements — and passed to a Claude Sonnet 4.5 planning agent.
Claude Sonnet generates a complete, execution-ready project plan: a Google Drive folder structure (client folder plus 5 named subfolders), a ClickUp board with 3 lists and all tasks with due dates and priorities, a branded client welcome email, an internal project brief, and a Slack notification. A Code node overrides the AI's free-form folder name with a deterministic version built from structured fields — preventing duplicate Drive folders on re-runs.
Drive folder creation and ClickUp board setup run in parallel. Idempotency is enforced at every step: the Drive search uses 'Always Output Data' with a downstream ID filter so the empty placeholder never triggers a false create; the ClickUp prep node returns an empty array if the folder already exists, stopping the task-creation chain cleanly. Client email, internal brief, and Slack alert fire last — the brief and Slack always send even if no client email address was found.