Cost per lead reduced $2 from $15 at 20,000 calls a day

"It used to cost us about $15 to create a lead in our largest vertical. That's come down to between $2 and $4 a lead. And we had at least 15 to 18% that were bad leads — we used to audit all of them with humans. With AI, that's come down to about 10%."
‍
- Mike Wedeking, Founder and CEO, Mile High Ltd

Mile High Ltd put a conversational voice agent on every lead-qualification call, cut cost per lead from $15 to $2–4, and let six people handle what used to take a team of 25.

Use case: Outbound insurance sales and lead qualification — every qualification call runs through the voice AI agent; human agents only take the qualified transfers

RESULTS AT A GLANCE

• 20,000 calls a day now run through the operation
• 100% of lead qualification is handled by the SigmaMind voice agent — humans no longer sit on the front of the funnel
• Headcount on qualification went from 25 agents to 6, and those six now work only qualified leads
• Cost per lead in the largest vertical fell from about $15 to between $2 and $4 — a 75–85% reduction
• Bad-lead rate dropped from 15–18% to about 10%
• Lead auditing moved off humans — every lead used to be reviewed by a person
• Onboarding was self-serve: create a username and password and start building

THE LIMIT WASN'T DEMAND. IT WAS HEADCOUNT.

Mile High Lead Services started as a lead-generation company out of Denver and grew into generating leads and running live call transfers for large insurance clients. The product was simple: qualify a prospect on the phone, then transfer the live call to a buyer. The bottleneck was also simple. Qualification sat on a team of 25 human agents, and that team was what capped how far the business could scale.

The unit economics underneath that cap weren't cheap either. In the largest vertical, a lead cost about $15 to create. And 15 to 18% of everything they produced was, by their own definition, a bad lead — which they only knew because they were auditing every single lead with human reviewers.

Neither number was the model of a company that wanted to run tens of thousands of calls a day.

WHAT THE AGENT ACTUALLY DOES

The voice agent now sits in front of the entire qualification process. Insurance sales and lead qualification are the primary use cases. Every inbound or outbound qualification call goes to the AI first. If the prospect qualifies, the call transfers to a human. If they don't, the agent closes it out without burning a seat on the floor.

That inversion is the whole operating change. Previously, 25 people were doing the filtering. Now the agent does the filtering, and six human agents handle what makes it through — the qualified leads, the transfers, the conversations that actually convert.

The volume that setup unlocked is 20,000 calls a day. That is not a volume 25 people could have absorbed by working harder. Concurrency, for a voice agent, is a configuration setting. For a call center, it is a hiring plan.

WHAT CHANGED: THE ECONOMICS

The cost per lead in that same largest vertical now runs between $2 and $4.

That's the number Mike keeps coming back to -

"It used to cost us about $15 to create a lead in our largest vertical. And that's come down to between $2 and $4 a lead."

For a business whose margin is the spread between what a lead costs to make and what it sells for, a change of that size doesn't optimize the model — it changes what's worth doing. Verticals that were marginal at $15, with 25 people on the phones, are comfortable at $3 with an agent that doesn't sleep and six people on the back end.

WHAT CHANGED: THE QUALITY

The bad-lead rate went the right direction too, and the audit itself stopped being a payroll line.

"We had at least 15 to 18% that were what we would call a bad lead. We would audit all of our leads with humans. With AI, that's come down to about 10%."

Two things happened at once. Fewer bad leads reach the buyer — roughly a third fewer — and the work of finding them no longer requires a team of people listening after the fact. The agent qualifies on the call, so quality is enforced at the point of contact instead of caught downstream. The six people left on the floor spend their time on leads that have already cleared that bar.

"Overall, it's much cheaper, much less expensive to utilize an AI agent for sure."

ONBOARDING: A NO-BRAINER

For a founder-led team without a dedicated AI engineering group, time-to-first-agent is the real adoption barrier. Mile High didn't hit one.

"The onboarding process is a no-brainer. You create a username and password and it's very easy. So far so good — in fact, really good."

No procurement cycle, no implementation partner, no multi-week integration project before the first call went out. The constraint that used to be headcount is now how fast they want to turn the volume up.

THE PARTNERSHIP

Mike describes his largest vertical as the one he's "really testing right now," with the core of the book in insurance and additional work across the other verticals the company has always covered.

What he singles out isn't a feature. It's the willingness to keep changing the product under him.

"I'm very excited about your thrust, your innovation, and your willingness to make changes and so forth. It's really been unique."