$20,000 down to $3. Same output.

Quick question: What if you could cut a $20,000/month cost down to $3 — without losing quality?

Not by cutting corners. Not by working people harder. Just by using AI where it actually belongs.

‍

Back in 2022, our sales team had five lead generators spread across different offices and countries.

Their job wasn't just finding prospects who matched our ICP. They did deep research — who owns the company, what they build, team structure, public mentions. Everything that later helped close deals.

One of them was extraordinary. Persistent, creative, relentless. She worked a lead for two years. Two. Years.

When he finally signed, he told her: "Nobody's ever been this persistent. Now I know you'll be a good partner."

He became one of our longest clients.

‍

LinkedIn kept blocking her account. Not because she violated rules — but because she was so effective they couldn't believe a human could work that hard. She extracted everything the platform had to give.

That team cost us about $20K a month in salaries alone. Not counting CRM subscriptions, LinkedIn licenses, research tools.

And they were worth it. 90% of our deals came through LinkedIn when most agencies struggled to close anything there.

At conferences, other founders kept asking: "Does LinkedIn actually work for sales?" They were skeptical. Surprised when we said yes.

We explained our process openly. We didn't hide it. But knowing the theory isn't enough. The real work lives in the details — the instructions, the workflows, the analysis, the knowledge transfer inside the team.

That's what made it work. Not just automation. Not just outreach volume. Human creativity and persistence, backed by clear systems and structure.

‍

Then everything changed

‍

Now there's AI. And tools like Clay growing like crazy because everyone needs enriched lead data. I looked at Clay. Solid product. But for me? 95% of what it does, I don't need.

Email addresses? Don't need them. Phone numbers? Not for me. AI personalization engines? Already have my own approach. And the price for features I'll never use? Didn't make sense.

So I built my own simple system.

‍

An AI agent in Make that uses Perplexity to collect public information and ChatGPT to analyze, structure, score, and prepare outreach recommendations.

Every morning, it updates my Notion CRM with fresh, relevant information about people I want to talk to. My assistant doesn't spend hours researching anymore. Neither do I.

The agent even scores each lead based on fit criteria and available information. Total cost for a system that saves 8 hours a day for two people?

$3 a month.

‍

Where it actually matters

‍

I'm not chasing clients. I'm having conversations with fellow founders. If I see a place where I can help — with advice from my experience or by offering the Ops-On-Demand Sprint — I do.

But I never shove a pitch down anyone's throat. Because even just exchanging knowledge with other founders gives me insights for my product.

Still, for those conversations to be valuable for both sides, I need to understand who I'm talking to and what challenges they might be facing.

That's where the enrichment matters.

‍

Recently, I reached out to an agency owner. The agent had found a podcast where she talked about almost not starting her business — because her father had gone bankrupt and it scarred her deeply.

I referenced that in my video message. It landed. She responded immediately. She didn't become a client — she's solo, and my offer is too much for her stage. But the point stands: That level of insight would've taken me hours to find manually.

The agent found it in minutes.

‍

The thing nobody expected

‍

Two clients I'm working with — both in high-ticket, handpicked client spaces — asked me to include this AI agent in my sprint deliverables.

It's not part of the offer. But they saw it in action during diagnostics when I asked how they research leads and why they don't. I showed them mine. They wanted it.

And honestly? I'm happy to share. We're already working together.

‍

But here's what's interesting:

This is maybe the tenth time in my entrepreneurial career I've seen this pattern.

The best validation for a business idea or product? Solving your own problem.

Not theorizing. Not guessing what the market needs. Just building something that fixes your actual pain — and discovering others have the same pain.

‍

What actually works

‍

AI is incredible when used correctly. But "correctly" doesn't mean replacing humans everywhere.

It means:

AI handles data collection and enrichment.

Humans handle the conversation.

‍

That's the same secret we learned years ago in our sales team: relevance and humanity win deals. Not volume. Not templates.

Back then, we paid five professionals to do the research. Now an agent does it 24/7 for $3.

But the outreach? The relationship building? The nuanced understanding of whether this person is actually a fit?

That's still me. And that's exactly how it should be.

‍

Something to try this week:

‍

Track how much time you (or your team) spend researching leads before outreach.

Not guessing. Actually track it for three days.

Then ask yourself:

• How much of that is gathering public information?
• How much is actual thinking about approach and fit?

The first part? That's what AI should handle.

The second part? That's where you add value.

‍

What I'm curious about: Are you using AI to replace the boring research work — or are you still doing it manually because "that's how it's always been done"?

Hit reply and tell me. I'm genuinely interested in what's stopping people from automating the obvious stuff.

‍

P.S. That lead generator who got blocked by LinkedIn multiple times? She's now running her own agency. Still just as relentless. Some things AI will never replace.