How a Series B SaaS in our portfolio generated 7,400 qualified leads in 90 days

The client had a senior six-person sales team, two SDRs, and an aggressive pipeline target for the back half of the year. Their existing outbound stack (Apollo + Lemlist + manual lists) was producing 200–300 qualified contacts per week, with reply rates in the single digits. Hiring more SDRs was out of budget. Without a new outbound motion, the pipeline target was at risk by mid-Q3.

We built a four-tier AI outbound stack: sourcing (Apollo + Vector intent data + scraped job-posting signals), enrichment (Clay waterfall of six providers), personalization (Claude with a three-tier budget), and reply handling (an AI triage agent that drafts replies, escalates hot leads to AEs, and books meetings via Calendly). Deliverability was hardened with a dedicated sending domain, a 14-day warmup, and continuous monitoring.

1. 01Sourcing: Apollo + Vector intent data + scraped job-posting signals; LLM ICP scoring assigns 0.0–1.0 fit
2. 02Enrichment: Clay waterfall (Apollo → Clearbit → Hunter → AI inference) reached 94% match rate
3. 03Personalization: tiered prompt strategy — top 50 prospects a week got five-line bespoke opens; next 200 LLM-personalized, rest templated
4. 04Multi-channel: email primary, LinkedIn DM secondary on score ≥ 0.8
5. 05Reply handling: AI agent classifies (positive / neutral / negative / OOO), drafts a reply, escalates hot to AE
6. 06Scheduling: Calendly auto-booking on positive replies, with prep notes attached
7. 07Eval harness: 52 golden examples re-run weekly to detect regressions

Within 90 days of going live, the client saw 7,400 qualified contacts processed, $2.1M in sourced pipeline, and a 38% reply rate. The AI SDR replaced the work of 2.5 SDRs without replacing any team members — instead, the SDR team moved up the stack to handle the warm replies and complex sales motions.

• Apollo
• Clay
• Vector
• Claude
• Smartlead
• Calendly
• HubSpot
• Slack
• Helicone