Client Overview
The Problem
The agency's three SDRs were each sending 500 emails a month from a shared bank of templates. Same opening line, same pain points, same CTA — just with the prospect's first name swapped in. Response rates had been declining for two years straight, and the sales team was demoralized.
The Head of Business Development had tried more templates, more A/B testing, better subject lines. Nothing moved the needle past 3%. The fundamental issue was clear: prospects can tell when they're getting a mass email. And they ignore it.
What they needed was real personalization at scale — emails that referenced something specific about each prospect's business, current challenges, or recent news. The kind of research that takes 20 minutes per prospect manually. Completely impossible at 500 emails/month per rep.
Core Pain Points
- ✗ 2% reply rate — 98 out of 100 emails completely ignored
- ✗ SDRs spending 60% of their time on manual research and writing
- ✗ Generic templates getting flagged by spam filters — deliverability declining
- ✗ Pipeline entirely dependent on referrals and inbound — outbound producing near-zero ROI
- ✗ No scalable way to personalize at the volume needed to fill pipeline
Our Solution
We built a fully automated prospecting and personalization engine. The system takes a prospect's name and company, scrapes their website and LinkedIn profile, extracts relevant signals (recent funding, new hires, pain points in job listings, tech stack, recent content), and passes that intelligence to Gemini API to write a unique, hyper-relevant email for each contact.
Every email sounds like it was written by a human who spent 20 minutes researching the prospect. Because essentially — the AI did. The SDRs now review and send. Research and writing: eliminated.
Prospect Intelligence Scraper
Custom scraper pulls data from prospect websites (about pages, blog posts, product pages), LinkedIn profiles (role changes, recent posts, connections), and news mentions. Outputs a structured prospect profile for each contact.
AI Email Generation via Gemini
Prospect profiles feed into a Gemini API prompt engineered to write concise, conversational cold emails (under 120 words). Each email references a specific, real detail about the prospect. No two emails are the same.
SDR Review Dashboard
Generated emails surface in a lightweight web dashboard where SDRs review, optionally edit, and approve with one click. Average review time: 45 seconds per email. Batch approval for high-confidence sends.
SMTP Delivery & Sequence Integration
Approved emails push directly to the agency's existing SMTP sending infrastructure with randomized send timing, follow-up sequence triggers, and reply detection that auto-pauses sequences when a prospect responds.
Tech Stack
Project Timeline
Discovery & Signal Mapping
Audited existing email sequences, identified what personalization signals actually drive replies, mapped data sources available per prospect type, and defined the ideal-customer profile for scraping targets.
Scraper + AI Pipeline Build
Built the scraping layer (website + LinkedIn), structured the prospect profile schema, engineered Gemini prompts through 200+ test iterations until email quality matched or exceeded manual writing, and integrated SMTP delivery.
SDR Dashboard & Testing
Built and deployed the review dashboard, ran a 50-email blind test comparing AI-generated emails vs. the agency's existing best templates. AI emails showed 3× higher reply rate in blind testing.
SDR Training & Full Launch
Trained all three SDRs on the dashboard workflow (30-minute session). Full production launch. First replies came in within 48 hours. 8 new client conversations booked within the first 30 days.
Results
Up from ~18% with generic templates. Subject line personalization and cleaner domain reputation both contributed.
Up from 2%. A 6× improvement — driven entirely by hyper-personalized first lines that referenced specific, researched details about each prospect.
Prospects converted to discovery calls within the first 30 days post-launch — from a channel that had generated zero the previous quarter.
Each SDR reclaimed ~3 hours/day previously spent on manual prospect research and email drafting — redirected to follow-up calls and deal management.
"I've seen a lot of 'AI email tools' that just spin up variations on the same template. This is completely different. The system actually reads a prospect's website and writes something that makes them think you spent an hour on it. We went from dreading cold outreach to it being our best-performing channel in under a month."
Ready to transform your outbound?
We'll audit your cold email workflow and show you exactly what AI personalization can unlock — free, no commitment.
Book a Free Strategy Call →