Project type: Portfolio project built end-to-end as a practical case based on a typical lead handling scenario. Not a client engagement. Tested with sample form submissions.

Context

A typical scenario: a company receives website inquiries through a contact form (for example, 80-150 per week). Each inquiry is reviewed manually before it is entered into internal systems. Marketing and sales need new leads organized by purchase intent, so that follow-up can be personalized and timely.

Problem

Manual lead handling in this setup typically causes:

• Time spent on routine work: reviewing inquiries and copying data into the CRM by hand.
• Slow speed-to-lead: inquiries wait until someone processes them, so high-intent leads can lose interest.
• Inconsistent segmentation: manual evaluation is subjective, which leads to uneven targeting in marketing campaigns.
• Limited scalability: processing effort grows with lead volume.

Solution

Built an integrated AI-driven workflow that replaces manual lead processing with instant real-time data distribution in CRM system:

• AI Purchase Intent Classification: Automatically extracts customer inquiry details and qualifies purchase intent (Hot / Warm / Cold) using LLM reasoning.
• Instant CRM & System Synchronization: Creates new leads in Zoho CRM, logs entries in Google Sheets, and routes contacts to targeted Klaviyo marketing segments in under 10 seconds.
• Real-Time Team Notifications: Triggers instant Telegram alerts for new high-intent leads, enabling sales reps to respond immediately.

Іmplementation Process

1. Prompt design

• Designed a structured classification prompt to evaluate customer comments and classify leads into Hot, Warm, or Cold categories.
• Implemented a strict JSON output schema with automatic fallback validation to ensure reliable data parsing.

2. API integration

• Connected Webhooks to capture website form submissions in real time.