Building an AI Sales Automation Pipeline That Sends 1,500+ Touches Monthly

The Problem

Manual outreach doesn't scale. When you're trying to reach hundreds of prospects across WhatsApp, email, and LinkedIn simultaneously, doing it by hand means you're either burning hours or dropping leads.

I needed a system that could:

• Discover and qualify leads automatically
• Enrich contact data from multiple sources
• Send personalized messages across channels
• Track responses and pipeline value

The Approach

I built Pulse -a multi-channel AI sales system running on a Contabo VPS. Here's the stack:

Lead Discovery

Using Exa.js for intelligent web search, the system finds prospects matching specific ICP criteria. It pulls company data, decision-maker names, and contact info automatically.

AI Personalization

Each message is generated by Claude API (Anthropic), tailored to the prospect's industry, role, and company context. No generic templates -every touch feels hand-written.

Multi-Channel Delivery

• Email: Sent through self-hosted Mailcow servers with proper DKIM/DMARC for maximum deliverability
• WhatsApp: Automated via Puppeteer browser automation
• LinkedIn: Connection requests with personalized notes

Pipeline Management

All interactions feed into a centralized dashboard tracking:

• Open rates and response rates per channel
• Pipeline value (₦10-15M quarterly)
• Lead scoring and follow-up scheduling

The Results

• 1,500+ monthly outreach touches across all channels
• ₦15M+ quarterly pipeline managed through the system
• Zero spam complaints thanks to proper email infrastructure
• 3x faster than manual prospecting

Key Takeaway

The secret isn't just AI -it's the infrastructure. Self-hosted email servers, proper domain reputation management, and intelligent rate limiting make the difference between landing in inbox vs. spam.

If you're building outreach automation, invest in your email infrastructure first. The AI personalization is the easy part.