🏢 Business / Industry: Marketing / Lead generation agency

🎯 Main Goal: Automate lead scoring and personalized follow-up based on AI qualification tier.

🔹 Context

Sales teams often spend significant time manually reviewing and following up with every incoming lead, regardless of quality. This creates inefficiency, slows response times, and dilutes focus away from the most promising prospects.

🔹 Problem

Too many unqualified leads were entering the pipeline and consuming valuable sales time. Without automated filtering, reps had no way to prioritize outreach — resulting in slow follow-ups, missed opportunities, and wasted effort on low-intent prospects.

🔹 Solution

An AI-powered lead qualification agent was built to automatically receive, score, and route every incoming lead — then trigger personalized follow-up messages via Klaviyo based on qualification tier (Hot, Warm, or Cold) — all without manual intervention.

🔹 Process

• Data ingestion: A webhook captures incoming leads and a data transformation step normalizes and maps fields to a consistent schema for downstream processing.
• CRM & logging: Each lead is simultaneously appended to Google Sheets and created in the CRM, ensuring full data traceability from the first touchpoint.
• AI qualification: A Basic LLM Chain powered by the Anthropic Claude model evaluates lead data against qualification criteria. A Structured Output Parser extracts a clean JSON score with tier classification (Hot / Warm / Cold).
• Routing logic: A Switch node routes each lead to the appropriate follow-up branch based on AI-assigned tier, enabling distinct handling strategies per segment.
• Automated follow-up: Each branch triggers a personalized email sequence via the Klaviyo API, delivering tier-specific messaging — high-touch for Hot leads, nurture flows for Warm, and re-engagement for Cold — with no manual input required.
• Notifications: A Telegram message is dispatched as a real-time alert to the sales team whenever a new lead enters the pipeline.

🔹 Tech Stack