Back to Cookbook

Forecast Modeler

Data-driven forecasts that replace gut feel

Forecasting shouldn't be rep opinions averaged in a spreadsheet. This skill builds forecasts from engagement signals, historical patterns, and stage conversion rates — producing confidence intervals, not guesses.

House RecipeWork5 min setup

INGREDIENTS

📄Google Docs💬Slack

PROMPT

Create a skill called "Forecast Modeler". Analyze my current pipeline and generate a forecast using: historical stage conversion rates, deal velocity by segment, activity-based engagement scoring, and close date reliability (how often do deals close on the date they say?). Output: commit number (80%+ probability), best case (50%+ probability), and upside (25%+ probability) with confidence intervals. Flag the top 5 deals most likely to slip with specific reasons. Show gap-to-quota: what additional pipeline or conversion improvement is needed to hit [quota]. Support scenario modeling: let me toggle deals in/out and see forecast impact. Track forecast accuracy over time: predicted vs. actual, by category.

How It Works

The skill analyzes your pipeline using activity data, historical conversion rates,

deal velocity, and engagement patterns. Instead of asking reps "what will you close?",

it calculates probability-weighted outcomes with confidence intervals.

What You Get

  • AI-generated forecast: best case, commit, worst case with confidence intervals
  • Historical conversion rates by stage, segment, and rep
  • Deals most likely to slip (with reasons)
  • Gap-to-quota analysis: what needs to happen to hit the number
  • Scenario modeling: "what if we close these 3 deals?"
  • Forecast accuracy tracking over time (predicted vs. actual)

Setup Steps

  1. Connect CRM with 2+ quarters of historical data
  2. Define your forecast categories (commit, best case, upside)
  3. Run initial calibration against last quarter's actuals
  4. Set weekly forecast delivery schedule

Tips

  • The model improves with more historical data — start even if imperfect
  • Compare AI forecast vs. rep forecast weekly to identify systematic biases
  • Use gap-to-quota analysis in pipeline reviews — it shifts the conversation from "what will close?" to "what needs to happen?"
  • Forecast accuracy tracking builds leadership confidence over time
Tags:#forecasting#pipeline#analytics#sales