The Problem
A multi-brand auto dealership had 60+ vehicles in stock at any time, listed across PakWheels, OLX, their own website, Facebook Marketplace, and a couple of WhatsApp broadcast lists. Each listing was maintained separately. When a unit sold, the team often forgot to remove it from one or two channels. leading to leads on already-sold cars and frustrated customers.
Financing applications were a nightmare too. Customers brought in scanned salary slips, bank statements, and CNIC images; an admin manually transcribed details into the bank's portal. taking 35-50 minutes per application with frequent typos.
What I Built
A three-part stack:
1. Single source of truth + multi-channel sync. All vehicles live in one Google Sheet that the sales team updates. An n8n workflow pushes changes to PakWheels (via API), OLX (via API), the website (Webflow CMS), and Facebook Marketplace (via Meta Graph API). When a car sells, marking it "sold" once removes it from every channel within 5 minutes.
2. Lead intake + distribution. Inbound enquiries from all channels (WhatsApp, calls transcribed by Whisper, web forms, IG DMs) flow into a single n8n queue. Claude classifies the enquiry by vehicle type, budget, and urgency, then assigns to the right salesperson and sends an auto-acknowledgement WhatsApp within 90 seconds.
3. Financing application AI. Customers upload their salary slip, bank statement, and CNIC photo via WhatsApp. A document AI pipeline extracts every field (name, CNIC number, employer, monthly income, bank, account number, etc.), validates against expected formats, and pre-fills the bank's portal. leaving only signatures for the admin to confirm. Application processing time: 35-50 min → 6 min.
Outcome
The dealership closed 2.3× the volume in 9 months without adding sales staff. The financing manager went from frustrated to evangelist. she now insists on the AI pre-fill before any application leaves her desk. The owner used the automation savings to launch a small subprime financing arm; the document AI pipeline handles those applications too.