Monitor HubSpot deal risk with OpenAI scoring and Slack alerts

How it works

This workflow runs on a daily schedule to analyze all active HubSpot deals and their latest engagement activity. It applies AI-driven behavioral scoring to predict conversion probability and deal health. High-risk or stalled deals automatically trigger Slack alerts. All insights are logged in Google Sheets for forecasting and performance tracking.

Step-by-step

• Step 1 – Trigger and collect active deals

  • Schedule Trigger – Runs the workflow automatically at a fixed time each day.
  • Get Active Deals from HubSpot – Retrieves all non-closed deals with key properties like value, stage, and activity dates.
  • Formatting Data – Cleans and normalizes deal data while calculating metrics such as deal age and inactivity duration.

• Step 2 – Enrich deals with engagement data

  • If – Filters only active deals to ensure closed deals are excluded.
  • Loop Over Items – Processes each deal individually to handle enrichment safely.
  • HTTP Request – Fetches engagement associations linked to each deal.
  • Get an engagement – Retrieves detailed engagement records from HubSpot.
  • Extracts Data – Structures engagement content, timestamps, and internal notes for AI analysis.

• Step 3 – Analyze risk and notify the team

  • AI Agent – Analyzes behavioral signals and predicts conversion probability, risk level, and next actions.
  • Format Data – Parses the AI output into structured fields and risk indicators.
  • Filter Alerts Needed – Identifies deals that require immediate attention.
  • Send Slack Alert – Sends a detailed alert with risks, signals, and recommended actions.
  • Append or update row in sheet – Stores analysis results in Google Sheets for tracking and forecasting.

Why use this?

• Detect deal risk early using consistent, AI-based analysis
• Reduce manual pipeline reviews for sales managers
• Provide clear, actionable next steps to sales reps
• Keep a historical log of deal health and forecasts
• Improve close rates through timely, data-driven intervention