How We Reduced Inquiry Response Time by 98% with AI Sales Automation

When a potential car buyer sends an inquiry, every minute matters. Research shows that responding within five minutes makes you 21 times more likely to qualify a lead than waiting just thirty minutes. For AutoFrance 3000, the official Bulgarian dealer for OMODA and JAECOO vehicles, inquiries were taking hours to process - and that delay was costing them deals.

We built an 80-node automation pipeline that cut their response time from hours to under two minutes, eliminated manual errors, and freed the sales team to focus on what they do best: closing deals.

Project: AutoFrance 3000

The Challenge

AutoFrance 3000 represents two rapidly growing Chinese automotive brands in the Bulgarian market: OMODA and JAECOO. With 6 vehicle models, each with multiple trim packages and pricing tiers, the sales team was drowning in manual work. Every single inquiry followed the same exhausting pattern.

A customer email would arrive. A sales representative would read it, try to identify which vehicle the customer was asking about, manually look up the correct specifications and pricing for that specific model and trim, write a personalised offer in Bulgarian, format it into a presentable PDF document, and send it back. Meanwhile, the lead data went into a spreadsheet - if the rep remembered to update it.

The problems were compounding:

• Response times averaged several hours, sometimes stretching into the next business day
• Pricing errors crept in when reps manually copied figures across trim packages
• Lead tracking was inconsistent - some inquiries fell through the cracks entirely
• Test drive bookings required back-and-forth emails with no calendar visibility
• The sales team spent more time on administration than actual selling

AutoFrance 3000 needed a system that could handle the entire inquiry-to-offer pipeline automatically, while still giving the team control over the final output. They were not looking to replace their salespeople - they wanted to remove every bottleneck standing between an inquiry and a professional response.

Our Solution

We designed and built an end-to-end sales automation pipeline using n8n as the orchestration engine, with 80 interconnected nodes handling every step from email arrival to CRM update. The architecture breaks down into five core stages.

1. Intelligent Email Monitoring and Classification

The pipeline begins with IMAP email monitoring that watches the sales inbox in real time. Every incoming message is immediately passed to GPT-4o-mini for classification. The AI reads the full email body and categorises it into one of three buckets: new inquiry, reply to an existing conversation, or irrelevant message. For new inquiries, the AI simultaneously extracts the customer's name, phone number, email, and the vehicle model they are interested in. This classification step alone eliminated the hours reps spent triaging their inbox each morning.

2. Automated Pricing and Offer Generation

Once a new inquiry is classified and parsed, the system matches the customer's vehicle interest against the live pricing database. Here is where the logic branches intelligently. For models with a single trim package, the system skips straight to generating a complete offer and sends it automatically - no human intervention needed. For models with multiple trim packages, it sends the customer a professional email asking them to select their preferred configuration, then waits for their reply before generating the final proposal.

3. Professional PDF Proposals

Every offer is rendered as a polished PDF document featuring the vehicle image, full specifications, pricing breakdown, and dealership branding. These are not generic templates - each proposal is dynamically generated with the customer's name, the exact model and trim they selected, and current pricing. The PDF generation runs entirely within the automation pipeline, producing dealer-quality documents in seconds.

4. Review Dashboard

While the automation handles the heavy lifting, the sales team maintains full visibility and control through a custom Next.js dashboard. Every processed inquiry appears in a clean interface where reps can review the AI's classification, see the generated offer, edit it if needed, and manually send it for edge cases. The dashboard also provides a real-time view of pipeline status - which inquiries are pending, which offers have been sent, and which leads need follow-up.

5. Calendar and CRM Integration

When a customer requests a test drive, the system creates a Google Calendar event with all relevant details and automatically updates the CRM record. Every touchpoint - from initial inquiry to test drive booking to follow-up - is tracked in a single timeline, giving the sales manager complete visibility into the pipeline without anyone manually updating a spreadsheet.

The Results

The impact was immediate and measurable. Within the first week of deployment, the numbers told the story.

• 98% reduction in response time: Inquiries that previously took hours to process now receive a professional offer in under 2 minutes
• Automated PDF proposals: Every offer is generated as a branded, professional document with zero manual formatting
• AI-powered email triage: GPT-4o-mini accurately classifies and routes every incoming message, eliminating inbox overload
• End-to-end pipeline tracking: Every inquiry is logged, tracked, and visible in a single dashboard from first contact to close
• Test drive calendar sync: Google Calendar integration with automatic CRM updates eliminated scheduling friction
• Sales team refocused: With administration automated, the team now spends their time on consultations and closing

The sales team went from spending the majority of their day on administrative tasks to spending it on high-value conversations with qualified leads. The automation did not replace them - it removed every obstacle between a customer's question and a professional answer.

Technology Stack

The choice of n8n as the automation backbone was deliberate. With 80 nodes spanning email parsing, AI classification, conditional branching, database lookups, PDF rendering, and multi-channel notifications, the pipeline needed an engine that could handle complex logic without vendor lock-in. Self-hosting also ensured that sensitive customer data and pricing information stayed entirely within AutoFrance 3000's infrastructure.