Context The client is a company with an active sales department managing dozens of concurrent deals inside a CRM. Sales managers and executives lacked real-time visibility into pipeline risks, and stalled deals were often noticed too late. Tracking the actual "health" of each deal and forecasting revenue required hours of manual data aggregation across calls, notes, and tasks.

Problem

Solution Designed and built a multi-tier Sales Intelligence automation system using n8n and an AI Agent. The system autonomously analyzes CRM data to generate daily funnel reports and proactively evaluates weekly deal health (scoring 1-10) based on accumulated CRM activities.

Знімок екрана 2026-05-28 203303.png

Process

  1. AI Sales Reporting: Built a daily workflow where an AI Agent analyzes the sales funnel and autonomously uses a tool to fetch deep details on problematic deals.

    Знімок екрана 2026-05-28 204109.png

Знімок екрана 2026-05-28 204122.png

  1. Sub-Workflow Architecture: Configured a weekly Deal Health workflow utilizing Loop Over Items and a dedicated Sub-workflow to aggregate notes, calls, and tasks for each specific deal.

Знімок екрана 2026-05-28 203446.png

  1. AI Scoring: Implemented an AI Agent with a Structured Output Parser to evaluate aggregated data and generate a JSON response with a Deal Health Score, risk factors, and recommended actions.
  2. Database Sync: Integrated Google Sheets with Append or Update logic (mapped by Deal ID) to maintain a historical database of deal health.
  3. Smart Routing: Set up a Switch node logic to instantly route "Critical" (Score ≤ 4) and "Medium" (Score ≤ 7) alerts to management via Telegram.

Results

Знімок екрана 2026-05-28 204027.png