Qualification Scorer

Score and qualify leads using a multi-axis framework. Takes a lead or list of leads, evaluates them on four axes (Fit, Timing, Access, Intent), classifies them into tiers, and produces actionable next steps for each.

Tools Used

• Read — load docs/icp.md, docs/signals/*.md, docs/accounts/*.md, CSV files
• Write — output scored results to docs/pipeline/scored-leads.md
• Glob — discover existing docs that can enrich scoring

Steps

Step 1: Context Loading

Check for existing files that inform scoring. Run these searches:

Glob: docs/icp.md
Glob: docs/signals/*.md
Glob: docs/accounts/*.md

• If docs/icp.md exists, read it and extract ICP criteria (company size, industry, stage, budget range). Use these criteria to auto-score the Fit axis.
• If docs/signals/*.md files exist, read them and extract known trigger events, intent data, and timing indicators. Use these to inform Timing and Intent scoring.
• If docs/accounts/*.md files exist, read them and cross-reference any leads against known account briefs for enrichment.
• If NONE of these files exist, ask the user:

I don't see an ICP definition at docs/icp.md. Before I can score leads accurately, I need to understand your ideal customer. Tell me:

1. What company size do you target? (employee count or revenue range)
2. What industries or verticals?
3. What company stage? (seed, Series A, growth, enterprise)
4. What budget range per month are you selling into?

Store the user's answers as the working ICP definition for this scoring session.

Step 2: Lead Input

Ask the user how they want to provide leads:

How would you like to provide the leads to score?

A. Paste a list (company name, contact name, title — one per line) B. Single lead deep-dive (I'll ask you detailed questions about one prospect) C. Give me a CSV file path to read

Handle each input mode:

• Option A (list): Parse the pasted text. Extract company name, contact name, and title for each row. Tolerate messy formatting (comma-separated, tab-separated, or line-by-line).
• Option B (single lead): Ask follow-up questions about the lead: company name, contact name and title, company size, industry, any recent events you know of, how you found them, any engagement so far.
• Option C (CSV): Read the file at the given path. Map columns to company, contact, and title. If column names are ambiguous, show the header row and ask the user to confirm the mapping.

Step 3: Scoring Framework

For each lead, evaluate four axes. Score each axis 1 through 5.

Fit (1-5) — Does this company match the ICP?

If docs/icp.md was loaded, auto-score Fit by comparing the lead's known attributes against the ICP criteria. If information is missing, default to 3 and flag it for user review.

Timing (1-5) — Is now the right moment?

If docs/signals/*.md files were loaded, cross-reference leads against known signals to auto-score. Otherwise, default to 2 and flag for user input.

Access (1-5) — Can you reach the decision maker?

IMPORTANT: Do NOT auto-score Access. Default to 2 for all leads and explicitly ask the user to review. The user almost always knows more about their access than any data source can reveal.

Intent (1-5) — How much buying intent is visible?

If signal files were loaded, use them to inform Intent scoring. Otherwise, default to 2 and flag for user review.

Step 4: Composite Scoring

Calculate the composite score and assign a tier for each lead.

Score = Fit + Timing + Access + Intent (maximum 20)

Tier Classification:

For each lead, generate a one-line recommended action that accounts for their specific score distribution. Examples:

• Fit=5, Access=1 -> "Find a warm intro path before reaching out — this is a perfect-fit account stuck behind a cold wall."
• Fit=3, Timing=5 -> "Timing is urgent but fit is uncertain. Validate ICP match before investing outreach effort."
• Fit=5, Timing=5, Access=5, Intent=1 -> "Everything lines up except intent. Test with a highly relevant value-drop to see if they bite."

Step 5: Interactive Review

Present the scores to the user in a table. Then ask:

Here are the scores for your leads. Do these look right?

Pay special attention to Access and Intent — you often know things about these that I can't see from the data. Tell me if any scores should be adjusted.

Reply with adjustments like: "Acme Corp Access should be 4 — I have a mutual connection through our investor" or "All looks good."

If the user provides adjustments:

1. Update the scores for the specified leads.
2. Recalculate composite scores and tier classifications.
3. Regenerate recommended actions if the tier changed.
4. Show the updated table.

Repeat until the user confirms the scores are accurate.

Step 6: Output

Write the final scored results to docs/pipeline/scored-leads.md. Create the directory if it does not exist.

Use this format:

# Scored Leads

> Generated: {YYYY-MM-DD} | Leads scored: {count}

## Summary

| Tier | Count |
|------|-------|
| HOT  | {n}   |
| WARM | {n}   |
| COLD | {n}   |
| PARK | {n}   |

## All Leads

| Company | Contact | Title | Fit | Timing | Access | Intent | Score | Tier | Recommended Action |
|---------|---------|-------|-----|--------|--------|--------|-------|------|--------------------|
| {data}  | {data}  | {data}| {n} | {n}    | {n}    | {n}    | {n}   | {tier}| {action}          |

(Sorted by score, highest first.)

## HOT Lead Engagement Plans

### {Company Name} — Score: {n}/20

- **Why HOT:** {brief explanation of what makes this lead top-tier}
- **Channel:** {recommended outreach channel — email, LinkedIn, warm intro, phone}
- **Angle:** {what to lead with — the specific pain point, trigger event, or connection to reference}
- **Timing:** {when to reach out and why — e.g., "This week. Their contract with X