AI Stack for Demand Gen Forecasting

Demand gen forecasts usually fail in one of two ways: the assumptions are invisible so finance can't stress-test them, or they're built from industry benchmarks instead of your own data. Both produce forecasts that fall apart in QBR. This stack builds a forecast grounded in what's actually true for your business — historical pipeline by channel, your own conversion rates, and the seasonality patterns in your own data. Claude generates a structured forecast framework with every assumption labeled so finance can model different scenarios. The Google Sheets model is yours to adjust; the Slack summary keeps leadership aligned without a formal presentation.

The Stack

The Prompt

This stack is built around the Exec Summary Generator Prompt. Here's the abbreviated version — the full prompt with all variables and usage notes is on its own page.

You are a B2B demand gen strategist building a quarterly forecast.

Using the historical pipeline data, channel performance benchmarks, and seasonality
patterns below, generate a structured forecast framework with:

1. Baseline pipeline projection by channel (using historical actuals, not benchmarks)
2. Key assumptions explicitly labeled: close rate, ASP, MQL-to-SQL rate by channel
3. Seasonality adjustments with rationale (which quarters historically over/underperform)
4. Upside and downside scenarios with the single variable that drives each
5. A Google Sheets formula framework for modeling each scenario
6. A 4-sentence Slack summary for leadership

The Workflow

1. Pull historical pipeline data by channel

   Export pipeline created per quarter for the last 4–6 quarters, broken down by channel, program type, and segment. Use actuals from your CRM — not what was forecasted. The gap between forecast and actual is often more useful than the numbers themselves.

2. Pull channel performance benchmarks from your own data

   Export cost per lead, MQL-to-SQL rate, SQL-to-opportunity rate, and opportunity-to-close rate by channel for the last 4 quarters. Use your own CRM data — industry benchmarks are a last resort when you have nothing else.

3. Document seasonality patterns in your pipeline data

   Which quarters consistently outperform or underperform your model? Are there fiscal year patterns, holiday effects, or end-of-quarter pipeline spikes that affect when deals actually close? Document what you observe, not what you expect.

4. Paste historical pipeline, channel benchmarks, and seasonality into the prompt

   Give Claude all three inputs together. Claude generates a forecast narrative and a model framework with every assumption explicitly labeled — close rate, ASP, MQL-to-SQL by channel — so finance can stress-test each variable independently.

5. Build the model in Sheets and distribute the Slack summary

   Move Claude's framework into Google Sheets. Adjust assumptions to hit your targets, validate the math with your finance partner, and send the Slack summary to leadership before your planning meeting. Alignment before the meeting beats defending a model in the room.

What This Replaces

• Demand gen forecasts built on spreadsheet math with no documented assumptions for finance to review
• Quarterly plans that don't account for historical seasonality patterns in your own pipeline data
• Forecasts that finance stress-tests by guessing at the methodology because it isn't documented

Related Stacks

• AI Stack for Demand Gen Planning →
• AI Stack for Pipeline Gap Analysis →
• AI Stack for Demand Gen Reporting →