n8n Workflow: Predict Salesforce Win Probability w/ Gmail + GPT-4o

Predict Salesforce win probability automatically—using Gmail engagement + GPT-4o

This n8n workflow watches for updates to your Salesforce Opportunities, analyzes recent Gmail activity tied to the deal, and uses OpenAI (GPT-4o-mini) to generate an updated win probability—then writes a refined, smoothed score back to Salesforce.

What this workflow does

• Triggers on a schedule whenever a Salesforce Opportunity is updated.
• Pulls full Opportunity context from Salesforce and derives key deal signals (including deal size, complexity, stage, and current probability metrics).
• Fetches recent Gmail messages and calculates engagement signals such as reply count and the days since the last activity.
• Calls OpenAI to predict win probability by sending both the engagement signals and deal context, receiving a JSON response that includes a new probability and a short reason.
• Validates and stabilizes the result by:
  • Checking whether the AI probability differs from the current Salesforce value by more than 5 points.
  • Clamping the output to a safe probability range.
  • Smoothing the updated probability against the current probability to avoid abrupt swings.
• Updates the Opportunity probability field in Salesforce with the refined value.

Use cases

• SaaS sales teams want data-driven win scoring that reflects both pipeline stage and real customer engagement.
• RevOps teams standardize opportunity scoring by combining Salesforce metrics with Gmail engagement signals.
• Automation engineers deploy a reliable Salesforce + Gmail + GPT n8n workflow with safeguards (thresholding and smoothing).

Technical details

• Trigger: Scheduled run on updated Salesforce Opportunities
• Integrations / nodes: Salesforce, Gmail, OpenAI (gpt-4o-mini), plus logic nodes such as if, set, code, and sticky note
• Field usage (Salesforce): Amount, Probability, Type, StageName, and the Opportunity probability update target