Cut my dialing seats from three to one, put a bot in front, cost per contract went down 40%

I've been trying to figure out which acquisition channel actually fits me and cold calling was winning on volume and losing on cost. Last spring I ran three VA dialers at $10.50/hr, about $5,450 a month fully loaded with data and dialer. I was signing roughly three contracts a month off it, so call it $1,800 per contract, which was survivable and not good.

In July I rebuilt it. One AI voice agent doing first touch on the list, no pitch, just confirming ownership and asking whether they'd consider an offer this year. Anything that answered yes or maybe got warm transferred or scheduled with my one remaining VA, who I promoted and paid $16/hr because she was now doing the only part that needed a person. Monthly cost landed at $2,900 including the voice vendor.

First month was bad. 1.4 contracts. The bot was ending calls on people who said something ambiguous, and the transfer window was 90 seconds so my VA missed live handoffs while she was on another call. Month two I widened the transfer to a callback queue instead of live-only, and rewrote the qualifying question to be one sentence. Went to 2.7 contracts, then 3.1, then 2.9. So roughly $1,000 per contract at about the same output.

What nearly broke it was the callback delay. A maybe that gets called back four hours later is worth much less than one you get in four minutes, and my VA could only cover eight hours. I ended up capping the bot's dialing window to her shift, which cut dials 40% and didn't cut contacts much at all, because I'd been dialing dead hours anyway.

What I'd keep: one strong human on the only conversation that matters. What I'd watch: the vendor's per-minute price, because my whole margin sits on it.

Answer

2paulgrant17PassivePRO5

The line I'd pull out is that cutting dials 40% barely moved contacts. That says your old three-seat setup was buying you volume you couldn't convert, so some of the 40% improvement isn't the bot at all, it's you stopping the waste.

On the vendor price, work out what happens at double. If per-minute goes from wherever it is to 2x, does $1,000 per contract become $1,300 or $1,900? If it's the second one, you've built a business on someone else's pricing page and they know it. I'd get the current terms in writing with a notice period, and I'd have a second vendor tested on a small slice so switching is a week and not a rebuild.

The other variable I'd poke is your one VA. She's now the whole conversion step at $16/hr. If she leaves, you're at one contract a month while you retrain, and the bot doesn't cover for her. What does month one after her notice look like on paper?

3 replies

2brandonlee36Passive

wren beat me to the single point of failure. kestrel, are you recording her calls and building a training set out of them, or is all the knowledge in her head?

1hannah_west29 Get yours

I want to argue with the $1,800 baseline. Three contracts a month across three seats reads like an average of a good quarter, and cold call output swings hard month to month. If your worst month on the old setup was 1.5 contracts and your best was 4.5, then comparing to a 2.9 average over three months is comparing a range to a point.

Run it again in six months with twelve months of both and I'd believe the 40%. The part I already believe is the labor mix, one $16 seat doing qualified conversations against three $10.50 seats reading a script. That's a better shape regardless of what the bot does, and it's the change I'd have made first even without the vendor.