Skip to main content

​
Qualify & Score Leads

To score leads against your ICP programmatically, use Deepline’s 2-step play: enrich each lead with firmographic data via enrich_company_finder, then classify with call_ai against your ICP criteria. Your scoring criteria live in the prompt, not in a template — so when your ICP evolves, you change the prompt instead of reconfiguring a template.
“Describe the goal and constraints, not the exact provider sequence.” Deepline selects the optimal enrichment path and applies your ICP criteria as a flexible AI classification — no template migration required.
B2B buyers are 57% through their purchase decision before engaging a sales rep (CEB/Gartner), so early, accurate qualification matters. This play scores prospects on company size, industry, funding stage, tech stack, and geography, then tiers them (A/B/C) so sales prioritizes the highest-fit accounts.

​
How do I score leads with Claude Code?

Tell Claude Code your ICP criteria and point it at your lead list. Deepline enriches each lead with real company data from 25+ providers, then uses AI classification to score against your criteria. No template configuration required.
“Score these leads against our ICP: B2B SaaS, 50-500 employees, Series A-C, US/EU. Tier them A/B/C.”
“Qualify leads.csv — our ideal customer uses AWS, has a security team of 5+, and revenue >$10M”
With Codex: 
codex "Score leads.csv against our ICP: B2B SaaS, 50-500 employees, Series A-C, US/EU. Tier A/B/C."

​
What does the lead scoring workflow do step by step?

Two steps: enrich each lead with real firmographic data from provider APIs, then apply AI classification against your ICP criteria. Every scored lead includes a tier, score breakdown, and written rationale.
1

Enrich company data

Each lead’s company is enriched with enrich_company_finder to get revenue, employee count, industry, funding stage, and tech stack. Waterfall enrichment across multiple providers delivers 20-40% higher data coverage than any single provider (Instantly).
2

Apply ICP criteria

call_ai evaluates each enriched lead against your stated ICP criteria, scoring on each dimension.
3

Tier and rank

Leads are classified into tiers (A/B/C or your custom scheme) with a rationale for each score.
4

Write scored output

Results are written with tier, score breakdown, and reasoning for each lead.

​
Which providers power the lead scoring?

Two tools run in sequence. Enrichment pulls real provider data (not AI-generated estimates), and classification uses AI to apply your custom criteria.
  1. enrich_company_finder — Gets structured company data (revenue, headcount, funding, tech stack) for each lead via a multi-provider waterfall
  2. call_ai — Applies your ICP criteria as a classification prompt, scoring each lead across every dimension you specify
You can score on any criteria: employee count ranges, specific technologies in the stack, geographic regions, funding stages, industry verticals, or custom signals like “recently opened a new office.” 
Score leads.csv against this ICP:
- B2B SaaS company
- 50-500 employees
- Series A through Series C
- US or EU headquarters
- Uses cloud infrastructure (AWS, GCP, Azure)

Tier A = matches 5/5 criteria
Tier B = matches 3-4
Tier C = matches 1-2

Write to scored-leads.csv with columns: company, tier, score, reasoning
Your ICP criteria live in the prompt, not in a template configuration. When your ICP changes, just change the prompt. No template migration needed.
Related: Enrich Company | Build Prospect List | Classify Company Signals

​
Frequently Asked Questions

​
How do I qualify leads against my ICP programmatically?

Tell Claude Code your ICP criteria in plain English: “Score leads.csv against our ICP: B2B SaaS, 50-500 employees, Series A-C.” Deepline enriches each lead with enrich_company_finder for real firmographic data, then uses call_ai to classify and tier each lead with a written rationale.

​
How do I update my scoring criteria?

Your ICP criteria live in the natural-language prompt, not in a fixed template. When your ICP evolves, change the prompt and rerun — no migration, no column remapping. This makes it easy to experiment with different scoring dimensions or adjust thresholds as your market understanding sharpens.

​
Can I score leads in bulk from a CSV?

Yes. Point Claude Code at any CSV with company names or domains. Each row goes through the enrich-then-classify pipeline. Output includes tier, breakdown, and reasoning per lead. Batches of 500+ work without configuration changes.

​
What data do I need to run lead scoring?

At minimum, company names or domains for each lead. enrich_company_finder handles the rest — pulling revenue, employee count, industry, funding stage, and tech stack from provider APIs. More specific ICP criteria yield more useful scoring output.

​
How much does lead scoring cost?

Each lead uses 1 enrich_company_finder credit and 1 call_ai credit. A batch of 100 leads costs approximately 100-200 credits depending on enrichment depth.