AI Stack for B2B Lead Scoring

Salesforce + Marketo → Claude → Salesforce Scoring Field + Slack
Matt Danese
Senior Demand Generation Manager. These stacks are built and used in production — not generated for a listicle.

Traditional lead scoring assigns points to actions — downloaded a PDF, visited a pricing page, clicked an email. It tells you someone is active. It doesn't tell you if they're a fit. This stack goes further. It pulls enriched lead records from Salesforce and Marketo, constructs a structured input with firmographic and behavioral data, and runs it through a Claude scoring prompt that evaluates ICP fit, intent signals, and disqualifiers together. The output: a scored lead with a confidence level and a reasoning summary that sales can actually use.

The Stack

Input
Salesforce Marketo
AI
Claude
Output
Salesforce (score field) Slack

The Prompt

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

Claude Prompt — Abbreviated
You are a B2B demand generation analyst evaluating inbound lead quality for a SaaS company.

Review the lead record below and score it against the ICP criteria provided.
Evaluate: company size fit, industry fit, job title/seniority fit, and behavioral intent signals.
Flag any hard disqualifiers that make this lead unsuitable for sales follow-up.
Output: ICP score (1–10), confidence level (High / Medium / Low), and a 2–3 sentence reasoning summary.
[ ... continued — see full prompt ]

The Workflow

  1. Pull enriched lead records

    Export new MQLs from Marketo or Salesforce — include firmographic fields (company size, industry, title), behavioral data (content consumed, pages visited), and any enrichment from Clearbit or Clay.

  2. Format the scoring input

    Paste each lead record into the Lead Quality Scoring prompt template. Include your ICP definition — target company size, industries, buying titles, and known disqualifiers.

  3. Run the scoring prompt

    Submit to Claude for each lead or batch. Claude evaluates ICP fit, intent signals, and disqualifiers — and outputs a score, confidence level, and short reasoning summary.

  4. Write scores back to CRM

    Update the lead record in Salesforce with the ICP score and reasoning note. Use a custom field so sales can sort leads by score before their daily outreach queue.

  5. Alert sales on high-confidence leads

    Post high-scoring leads (score ≥ 8, confidence: High) to a dedicated Slack channel with the reasoning summary and a link to the CRM record.

What This Replaces

Related Stacks

New stacks drop weekly.

Each one includes the tools, the Claude prompt, and the workflow logic. Free — built for in-house B2B demand gen managers.

← Back to Stack Library

© 2026 The Demand Engineer · thedemandengineer.com