There's a fundamental truth about automation that most people miss: the best system isn't the smartest one—it's the one that actually runs.
Over the past week, I've been watching my automated prospecting system quietly do its job. While I'm having coffee, it's finding dental practices in Fort Lauderdale. While I'm asleep, it's queueing up law firms in Palm Beach. No supervision required.
The Setup
The system runs on a simple loop:
- Cron triggers at set intervals throughout the day
- Web search finds businesses matching specific criteria (industry, location, keywords)
- Data extraction pulls contact info, phone numbers, websites
- Queue insertion adds qualified prospects to the right campaign
- Notification pings me with what it found
That's it. No ML models. No complex scoring algorithms. Just targeted searches, basic filtering, and consistent execution.
Multiple Campaigns, Same Engine
The real power comes from running multiple campaigns simultaneously:
- Receptionist campaign: Medical practices, dental offices—anyone fielding calls
- Reviews campaign: Businesses where reputation drives revenue
- AI campaign: Firms that would benefit from automation
- Boring campaign: Generic outreach to professional services
Same infrastructure, different targeting. Each campaign has its own search queries and qualification criteria, but they all feed into the same pipeline.
Why "Good Enough" Beats "Perfect"
Here's what I've learned: a mediocre system that runs constantly outperforms a perfect system you never ship.
My prospecting automation isn't sophisticated. It doesn't use AI to score leads. It doesn't do predictive analytics. It just finds businesses that match basic criteria and adds them to a queue.
But it runs 6+ times per day, every day. Over the past week, it's added 80+ prospects to various campaigns. While I was building other things. While I was traveling. While I was literally asleep.
The Technical Reality
The stack is intentionally simple:
- Cron jobs for scheduling (built into the platform)
- Brave Search API for finding businesses
- Web scraping for extracting contact details
- PostgreSQL for the prospect queue
- Webhook notifications to keep me informed
Total setup time was probably 2-3 hours. Most of that was tuning the search queries to find quality prospects instead of spam.
Lessons Learned
1. Rate limits are real. Brave Search will throttle you. Build in delays, handle failures gracefully, and don't try to be greedy.
2. Garbage in, garbage out. The quality of your search queries directly determines the quality of your prospects. Spend time here.
3. Separate finding from reaching out. The prospecting system just finds leads. Outreach is a separate process. This separation keeps both systems simpler.
4. Notifications matter. I get a ping every time the system adds prospects. It takes 5 seconds to review and I stay in the loop without babysitting.
The Math That Matters
Let's say manual prospecting takes 30 minutes to find 5 good leads. That's 6 minutes per lead.
My automation finds ~15 leads per day with zero effort. If I value my time at $100/hour, that's $150/day in saved time—over $4,000/month.
And the system gets better over time as I tune the queries based on what converts.
Ship It Ugly
If you're thinking about building automated prospecting, here's my advice: ship it ugly.
Don't wait until you have the perfect scoring algorithm. Don't build a beautiful dashboard first. Get something running that finds leads and adds them to a list.
You can make it smarter later. But you can't get back the leads you missed while you were perfecting your system.
The 80% solution that runs 24/7 beats the 100% solution that's still in development.
Start simple. Ship it. Let it run while you sleep.
Top comments (0)