📌 Project Overview

• Role: Automation & AI Systems Architect

• Platform: n8n (Self-Hosted/Cloud)

• Stack: Google Sheets API, OpenAI API, JavaScript Node, Slack API

• Key Outcome: Created an automated, self-cleaning lead-triage engine that instantly prioritizes high-value software enterprise leads while seamlessly handling lower-tier submissions.

• 📉 The Challenge

• ⚙️ The Technical Solution & Architecture

1. Smart Data Ingestion & State Control

• Implemented an n8n Google Sheets trigger that checks for incoming lines.
• Established a strict conditional rule filtering for Processed = No rows. This ensures that the workflow only acts on new data, preventing redundant operations and saving massive API execution costs.

2. Custom JavaScript Data Parsing Engine

• Integrated an LLM node to analyze raw company websites and profiles.
• Because the raw AI output returned unformatted markdown text, I wrote a custom JavaScript parsing engine to isolate specific tokens, map raw variables, and structure the data into highly predictable JSON object trees.

3. Dynamic Conditional Routing Switch

• Configured a conditional Switch node evaluating lead scores (Threshold: 70+).
• Hot Leads (Enterprise): Instantly builds custom Slack message text layouts and dispatches real-time alerts directly into the sales team's pipeline channel.
• Low-Priority Leads: Dynamically branches away from critical sales views, automatically logging records into a secondary Google Sheets nurture queue.

4. Closed-Loop Relational Database Patching

• To eliminate the threat of infinite execution loops, I built a terminal database update step.