I spent $457 on one AI outbound voice agent. It generated $6,800 in 15 days.
That number sounds great in a headline. But if you copy what I did without understanding why it worked, you'll almost certainly lose money. That's exactly what happens to most businesses that bolt AI onto a broken process.
So this isn't a hype video. It's the economics, and a decision framework for whether voice AI even belongs in your operation.
It started as a follow-up problem, not an AI project
We ran an ascension funnel: people bought a low-ticket course, then a percentage upgraded to high-ticket coaching. After a few years, we had thousands of buyers going cold in the CRM. Human follow-up doesn't scale cleanly, so those leads sat there.
The question was never "can AI make outbound calls". Technically, obviously it can. The real question was: can follow-up be systematized without damaging trust, brand, or qualification quality? Because at scale, bad follow-up doesn't just waste money. It burns reputation, which is the most expensive thing you own.
The economics, in full
Fifteen days. Every buyer on the list had purchased before, so it was re-engagement, not cold calling.
| Metric | Number |
|---|---|
| Spend | $457 |
| Cost per dial | $0.04 |
| Leads called | 2,825 |
| Total dials | 11,000+ |
| Cost per booked call | ~$20 |
| Deals closed | 2 ($4,800 + $2,000) |
| ROI | ~15x |
For contrast, our human setters cost anywhere from $60 to $200 per booked call, on top of a retainer around $3,000 a month. Setters are still worth it, more calls beat fewer. But the cost structure here is a different game. Full KPI breakdown is here.
Where AI actually fit
The agent did four narrow things, and no more: it called existing buyers, followed a script our team had already validated, qualified on BANT, and booked only when the lead met the bar. We didn't invent a process for AI. We copied one that already booked calls.
Three lessons to keep
- AI is leverage, not strategy. It's the multiplier you put on top of a strategy that already works, not the strategy itself.
- Outbound works best on existing demand. Buyers and opt-ins, not strangers.
- Consistency beats persuasion at scale. A team of solid 7-out-of-10 reps following one proven script beats two brilliant reps and a lot of average ones. The average is what scales.
The one rule
Here's the filter I use, and I'd use it on your business too. If a human can't run this process consistently, AI shouldn't touch it.
AI doesn't fix weak fundamentals. It exposes them faster. Weak systems don't get strong at scale, they fail at scale. So if your offer isn't proven, your script doesn't convert, or your leads aren't qualified, fix that first. Then AI can take what works and pour fuel on it.
-Ruben
Grow your revenue or cut your costs in 90 days. Or you don't pay.
Done-for-you AI systems, the same ones I run in my own agency and build for 7 and 8-figure coaching and consulting businesses. Want them installed for you? Let's talk.
