Case Study 03 · B2B Sales / SaaS
AI Cold Email Lead Qualification System
Every lead scored and personalized before a rep ever sees it.
Client
Confidential B2B Sales Team (NDA)
Industry
B2B Sales / SaaS
Stack
n8n, OpenAI, CRM, Email API, Enrichment API
Meaningfully higher
Response rate
Hours → Minutes
Time to qualification
More per rep, per week
Booked meetings
The business challenge
The sales team was generating a healthy volume of inbound and outbound leads, but reps spent most of their week manually researching each one, deciding whether they were worth pursuing, and writing individual outreach emails — before finding out most weren't a good fit at all.
The bottleneck wasn't lead volume, it was qualification. Reps' time is expensive to spend on research and generic outreach, and by the time a genuinely good-fit lead was identified, the personalized email that should have gone out immediately was often delayed behind everything else in the queue.
Objectives
- Automatically enrich every new lead with the firmographic data reps were manually looking up
- Score each lead against the team's actual ideal-customer profile, not a generic scoring model
- Generate a genuinely personalized first-touch email for qualified leads, not a mail-merge template
- Sync everything into the CRM with automated, scheduled follow-up for no-replies
The solution
New leads are enriched automatically the moment they enter the system, then scored by AI against the team's specific ideal-customer profile. Leads above the qualification threshold get a personalized outreach email generated from their enrichment data and sent automatically; everything is logged in the CRM with a scheduled, automated follow-up sequence for leads that don't respond.
System architecture
Leads flow through enrichment and AI scoring before hitting a qualification router: leads above threshold move into personalized email generation and CRM logging, while sub-threshold leads are logged for visibility without consuming rep time.
Workflow breakdown
Lead Capture
New leads enter from forms, list imports, or CRM triggers and are immediately queued for enrichment.
Enrichment
Firmographic and contact data (company size, industry, role, tech stack signals) is pulled automatically from an enrichment API.
AI Scoring
Each enriched lead is scored against the team's ideal-customer profile, with reasoning attached to the score.
Personalized Outreach
Leads above the qualification threshold get a first-touch email generated from their specific enrichment data — not a template with fields swapped in.
CRM Sync
Every lead, score, and email sent is logged in the CRM automatically, keeping reps' pipeline view current without manual entry.
Automated Follow-Up
Leads that don't respond within a set window automatically receive a scheduled follow-up sequence, with reps alerted the moment a reply comes in.
Technical implementation
- n8n orchestrates enrichment, scoring, and email generation as a single pipeline triggered the moment a new lead enters the system.
- An enrichment API supplies firmographic and role data, normalized into a consistent schema before being passed to the scoring step.
- The AI scoring step is prompted with the team's actual ideal-customer-profile criteria and returns both a numeric score and a short written rationale, visible to reps directly in the CRM.
- Qualified leads' enrichment data feeds directly into email generation, so each message references something specific and real about that company rather than generic personalization tokens.
AI prompt strategy
The scoring prompt is grounded in the team's actual closed-won and closed-lost deal patterns rather than generic firmographic rules, and it's required to return a written rationale alongside the score — improving scoring quality and giving reps something concrete to sanity-check. The email-generation prompt is deliberately restricted to only reference facts present in the enrichment data, keeping personalization genuine instead of invented.
Integrations used
Results
Reps stopped spending their week on research and generic outreach and started spending it on conversations with leads the system had already confirmed were worth their time. Qualification that used to take most of an hour per lead now happens automatically in minutes, and outreach goes out while the lead is still warm.
Lessons learned
- Grounding the AI scoring prompt in real closed-won/closed-lost patterns — instead of generic firmographic rules — produced scores the sales team actually trusted and acted on.
- Requiring a written rationale alongside every score mattered as much as the score itself, since it's what got reps to trust and adopt the system quickly.
- Restricting the email-generation prompt to only reference verified enrichment data was essential — early tests without that constraint occasionally 'personalized' with plausible-sounding but incorrect details.
Want a system like this for your business?
Tell me about your process and I'll get back to you within 24 hours with next steps.