Context

• Industry: B2B Services / AI Automation & Telegram Bot Development.
• Business: Capturing inbound leads from a company website (Wix) to acquire clients for custom automation workflows and bot development.
• Legacy Process: Incoming inquiries from the Wix contact form were manually processed by a sales manager (data verification, lead scoring, manual CRM data entry, and email follow-ups).

Problem

• Delayed Response Times (High Response Time): Hot leads waited anywhere from several hours to a full day for a response, significantly degrading conversion rates.
• Inefficient Lead Scoring: Managers spent equal time analyzing low-intent inquiries, spam, and high-value B2B project requests.
• Fragmented Data: Submission data was scattered across email inboxes, CRM records, and spreadsheets, creating analytical blind spots and hindering timely follow-ups.

Solution

An autonomous AI-driven system built on n8n and OpenAI API (GPT-4o) that operates in real time to:

1. Capture submission payloads via Wix Webhook and standardize the incoming dataset.
2. Perform deep AI analysis on user comments and contact details to classify leads instantly as HOT, WARM, or COLD.
3. Generate concise intent summaries, categorize project scope, and assign a numerical Lead Score (1 to 10).
4. Dynamically route enriched lead profiles to Klaviyo (for automated email flow triggers), Zoho CRM, Google Sheets, and issue instant push alerts to Telegram.

Process

1. Prompt Design (AI Classification Engine)

Engineered a specialized System Prompt with strict evaluation criteria that mandates valid JSON output:

• HOT (Score 8–10): Well-defined business requirements, explicit deadlines, direct requests for n8n/Telegram bot engineering, corporate contact info.
• WARM (Score 5–7): General pricing inquiries, requests for initial consultation, lack of clear technical specifications.
• COLD (Score 1–4): Irrelevant comments, test submissions, invalid contact info, vendor spam, or unsolicited service offers.