A Christchurch new-build developer was watching hot Meta leads go cold. Not because the ads weren't working. Because by the time the SDRs got to them, the buyers had already gone to look at the competitor down the road.
Here's how they fixed it in 14 days, with the numbers.
The client asked to stay anonymous for now, so we have stripped the company name and the specific show home addresses. The data is unchanged.
The Setup
The setup, and what was breaking
The team builds 2 to 3 bedroom townhouses in the southwest Christchurch catchments. Halswell. Rolleston. Wigram. The buyers are mostly first-home buyers on KiwiSaver, looking in the $1m to $1.3m range, browsing Facebook on the couch at 9pm. Meta lead-form ads have been the core acquisition channel for over a year.
Two SDRs handled the call-back queue. They were good. They were also drowning. A weekend's worth of leads would land Monday morning and most wouldn't get a call until Thursday afternoon, sometimes Friday. By then, half had already booked a viewing with another developer.
That was the gap they wanted to close. Not lead generation. Lead response time.
The Funnel
What 931 dialled Meta leads produced
In a 14 day pay-as-you-go trial, the AI voice agent called 931 Meta leads. Same day, often within minutes of the form hitting the CRM.
Here's what came out the other end.
| Outcome | Count | % of all dials |
|---|---|---|
| Humans answered | 348 | 37% |
| ↳ Qualified prospect | 71 | 7.6% |
| Viewing booked at show home | 49 | 5.3% |
| Tagged for 3 to 6 month follow-up | 22 | 2.4% |
The booking number is the one to focus on. 49 confirmed viewings at the show home, in 14 days, from one channel, with no SDR doing the cold legwork. 69% of every qualified prospect ended the call with a slot in the calendar.
Daily pipeline
14 days · 49 bookings total · two days did half the work
Day 8 (Tuesday) and Day 11 (Friday) are the spikes. 28 of the 49 bookings landed on those two days. Both followed a campaign refresh that landed fresh Meta leads, and both fell on the agent's strongest converting weekdays.
The Math
The math, $7.12 per booked viewing
$7.12 per qualified booked viewing
Total trial spend was $505.70 across 14 days. Divided across the 71 qualified prospects, that is $7.12 each. For a developer selling $1m+ townhouses, the maths doesn't need rounding.
- Total trial spend (14 days, pay-as-you-go)$505.70
- Cost per dial$0.54
- Cost per answered call$1.45
- Cost per qualified prospect$7.12
The other number that mattered to the team:
21+ hours of SDR time reclaimed
That's roughly 1.5 to 2 weeks of full-time calling the dead numbers, the voicemails, and the "wrong number" replies every Meta list throws up. The two SDRs got those hours back to do what only humans can do: walk a buyer through the floor plan, follow up after a viewing, close.
The wage comparison
Two SDRs at $30/hr
$2,400+
In wages alone for 80 hours of dialling. Before KiwiSaver, ACC, holiday pay, on-costs.
AI agent, flat trial cost
$505.70
Same dials. Same conversations. No on-costs. No PTO. No "they got pulled into a meeting."
When It Lands
When the bookings actually landed
Three different cuts of the same data. The patterns are the operator manual: when to call, what days to lean in, when to pause the agent.
Bookings by hour of day
All times NZT · max 10 in any hour
Hour of day (NZT)
11am peaks at 21% conversion. 10am to 1pm is the golden window. A second wind 4pm to 6pm, helped by buyers checking their phone after work. Calls outside 9am to 7pm are wasted.
Bookings by weekday
14 days · two days own the week
Day of week
Tuesday and Friday own the week. 31 of 49 bookings between them. Wednesday and Sunday are wastelands. Pause the agent on those days, run double-density Tue and Fri.
Conversations
What buyers said, and what the AI said back
We scanned every transcript for the patterns that came up most. Here's what buyers actually said, what the AI did, and how often it still booked the viewing.
Price & deposit
38%
Came up in 24 calls · 9 booked
Buyers asked "how much." The AI anchored the actual range, explained what moves the price (section size, finish package), then asked a question that sorted the buyer into the right tier. No dodging the number, no "let me put you through to someone."
Agent
"Our 3-bedroom homes in Halswell sit between $1.05m and $1.25m, depending on section size and the finish package. Bigger sections push you toward the top of that. Are you focused on the entry-end pricing, or are you open to the larger sections with the bigger backyards?"
Specs (bedrooms, schools, parking)
36%
Came up in 77 calls · 28 booked
Cold recall on the spec sheet. The AI answered the spec question with the full picture (parking, school zone, who the layout suits), then asked the deeper question that surfaces what the buyer actually cares about.
Agent
"Our 3-beds in Halswell come with internal single garages and are zoned for Knights Stream and Aidanfield Christian. The 4-bed plans add a study nook off the lounge, which works for households where someone runs a business from home. Is the school zone the priority for you, or is layout the bigger factor?"
KiwiSaver & first-home eligibility
35%
Came up in 85 calls · 30 booked
Buyers wanted to know if they qualified. The AI gave them the actual rules and dollar figures (caps, contribution period, First Home Grant on a new build), then asked the question that surfaces the real blocker, the bit that catches most people out.
Agent
"Quick rundown: three years in KiwiSaver, earning under $95k single or $150k as a couple, and you can pull most of your KiwiSaver out for the deposit. The First Home Grant adds $10k per person on a new build like ours, so $20k if you're buying as a couple. Have you confirmed your KiwiSaver balance and your three-year contribution record? That's usually the bit that catches people."
"Call me back" / busy
18%
Came up in 34 calls · 6 booked
Lead said "not now, I'm busy." Instead of pushing for an immediate slot, the AI offered a low-friction handoff (text with the price range and floor plans) and let the buyer pick the next step on their terms. Lower booking rate by design, but no leads burned by being pushy.
Agent
"All good, I'll keep it short. Want me to text you the price range and a link to the floor plans right now so you've got something to look at when you get a moment? Then if you'd like a proper chat, just reply with a time and we'll lock it in, after-hours is fine too."
13 calls hit "wrong number," "didn't sign up," or "not interested." The AI ended every one cleanly. Of the rest, plenty weren't pre-approved or were too early in their search to book today. The AI tagged those for an automated 3 or 6 month callback. 277 live conversations the two SDRs didn't have to take, so they could focus on closing the 71 prospects who were ready.
The Lever
Speed-to-lead, the biggest lever
We split the trial into two cohorts and ran them through the agent separately to see what speed actually does to conversion.
Same-day Meta leads
Called within minutes of the form hitting the CRM
- Calls527
- Bookings32
- Cost per booking$9.52
- Book rate of pickups15.6%
Cold backlog
3+ months stale, sitting unworked in the CRM
- Calls385
- Bookings13
- Cost per booking$13.78
- Book rate of pickups10.1%
31% cheaper. 55% better conversion.
Same-day Meta beat the cold backlog on every metric. Speed-to-lead is the single biggest lever in this whole trial. Faster pickup, fresher buyer, more booking. Every minute a Meta lead sits in a queue is conversion you're losing.
Honest Framing
The honest caveats
Three things to be straight about.
The Pipeline
What's next
On the same conversion rate, an extra month at the same daily volume would book another 100+ viewings. Even at half that rate, the maths still works.
The team has now wired the agent into the lead webhook directly so every Meta form submission triggers a call within minutes of a buyer hitting submit. The 22 "not ready, follow up in 3 to 6 months" prospects are already queued for the next campaign, with the agent opening from the context of the last conversation: "Last time we spoke you said you were saving until winter, just checking in on where you got to."
That's the difference between a one-off lead-response patch and a pipeline that keeps compounding. Same agent. Same patch. Just better timing every cycle.
Want to see what your numbers would look like
A 15 minute demo on your own list. We open the campaign builder, plug in your lead source, and show you what the funnel and the cost per booking would look like for your developments.
Book a demo · AI Voice Agents pillar · Sydney agent: 141 vendor leads in 90 days
Leonardo Garcia-Curtis
Founder & CEO at Waboom AI. Building voice AI agents that convert.
Ready to Build Your AI Voice Agent?
Let's discuss how Waboom AI can help automate your customer conversations.
Book a Free Demo
