By Oskar Korjus, co-founder of luup voice agents for real estate. The launch math below is built from public response-time benchmarks and luup's own build figures, not invented client outcomes.

Real estate presale launch lead response: surviving the inquiry spike

A presale launch is the most expensive 72 hours in a developer's calendar. Paid media is live. The agent network is activated. The sales gallery is staffed. And the registrations land all at once. This is the moment your real estate presale launch lead response either captures the pipeline or quietly leaks it. Most leak it.

Not because the sales team is weak. Because the volume is physically impossible to hold live. A human can run one conversation at a time. A launch sends 50 at once. The gap between those two numbers is your leak, and it is bigger than most developers ever measure.

1. The leak: what an unanswered launch inquiry actually costs

Start with the cost of inaction, because that is the number that moves budgets. A missed inquiry on a 300,000 dollar unit at a 3% commission is 9,000 dollars of gross revenue gone. That is one inquiry. One. Now hold that figure while the launch math compounds it. Existing-home sales run in the millions of units a year, per the National Association of Realtors, so the buyer pool is deep and impatient.

Speed is the whole game. Lead response that arrives within minutes converts at a multiple of response that arrives hours later, a pattern documented across decades of sales research summarized by Harvard Business Review. Buyers do not wait. They register with the first three projects that answer, and the slow developer never enters the consideration set. The average team takes hours to call back. By then the buyer has booked a visit somewhere else.

Then there is timing. Launch ads run in the evening, on the couch, when intent is highest and your gallery is closed. The buyer registers at 9pm. Nobody answers until tomorrow, and tomorrow the interest has cooled and the competing project has already booked the visit. Speed-to-lead is the single biggest controllable variable in real estate marketing, a point McKinsey makes repeatedly about high-consideration purchases.

So the leak is not one thing. It is three stacked failures: too much volume to answer live, too slow when you do answer, and nobody staffed when most inquiries actually land. Each one alone costs units. Together, at launch scale, they cost the launch.

2. Why presale demand arrives as a spike, not a stream

A presale system is not launch-and-leave. Done right, it runs 6-12 months aligned to construction milestones, with demand pulses at groundbreaking, structure topping-out, facade completion, and handover. The teaser phase builds a waitlist. Then launch fires.

And launch fires hard on purpose. The point of a launch is concentration. You stack paid media, broker activation, and the warmed waitlist into a narrow window to manufacture urgency and signal that units are moving. That is good marketing. It is also exactly what overwhelms the phones. The same compression that creates buyer urgency creates an inbound wall your team cannot scale to meet, because you cannot hire 20 agents for one weekend and fire them Monday.

This is the structural trap. The better your launch marketing performs, the worse your response capacity gap gets. A launch that underperforms never strains the team. A launch that hits its registration target buries them. Most developers only discover this on launch day, watching the registration counter climb while three phones ring unanswered. The ad-factory work that drives the spike and the response layer that catches it have to be built together, which is why we pair this with the ad factory for real estate project launches.

3. The 300-registration math, run cold

Numbers, not adjectives. Take a mid-market off-plan launch over one weekend.

Three people answering 40 live across a busy weekend is generous. Each real conversation runs 8 to 12 minutes, plus the qualifying, plus the gallery scheduling, plus the buyer who needs three call-backs. The team is not idle. They are maxed. And they still leave 260 registrations untouched.

Now price the 260. Assume a conservative 25% were genuinely qualified buyers, not tire-kickers. That is 65 real prospects. Apply the 9,000 dollar gross-commission value of a single 300,000 dollar unit, and even a modest conversion off those 65 prospects is six figures of pipeline left on the floor in 48 hours. You spent the paid budget to create those 260 registrations. Then you let most of them expire unanswered. That is the most expensive part of the launch, and it never shows up on a media report.

The fix is not more people. Three becomes five and you still cannot hold 260 simultaneous conversations after hours. The fix is parallelism: a layer that answers all of them at once, instantly, and routes the qualified ones to your humans. Run your own version of this on the revenue leak heatmap with your unit price and registration estimate.

4. The named system: a launch-surge voice layer

The system is a launch-surge voice layer. One job: absorb every simultaneous inbound during the 2-4 week launch window so zero registrations go dark. It runs in parallel, so 50 registrations in the same minute get 50 conversations in the same minute. No queue. No voicemail. No tomorrow. The voice itself is built on production infrastructure like Vapi for orchestration and natural speech models, not a hold-music phone tree.

It does four concrete things on every call:

• Answers at a 90-second SLA. Every inbound, day or night, gets a live conversation inside 90 seconds. This is the number that puts you in the first-responder bracket instead of the multi-hour average.
• Qualifies on three axes. Budget band, unit type, and purchase timeline. The buyer who wants a two-bed within 90 days gets tagged differently from the one browsing for next year.
• Books the sales-gallery appointment. Qualified buyers get a real slot in the gallery calendar before the call ends, while intent is at its peak.
• Tags hot leads for human follow-up. The agent does not pretend to close. It captures, qualifies, books, and hands the warm ones to your sales team with full context so the human call is a closing call, not a discovery call.

The design rule is simple. The voice layer does the volume work no human can scale to. Your people do the relationship work no machine should fake. That split is the whole point. We go deeper on the build in the AI voice agent guide for real estate developers, and the exact dialogue is in the inbound voice agent script template.

5. The call flow on a live launch night

Here is what one inbound looks like end to end, at 10:47pm on launch Saturday, with 40 other calls running in parallel.

5.1 Answer and frame

The buyer taps call from the launch ad. The agent picks up inside 90 seconds, names the project, and confirms they are registering interest. No hold music, no all-agents-are-busy script. The single biggest signal of a serious developer is that someone answered.

5.2 Qualify in under two minutes

Three questions, conversational, not a form read aloud. Budget band. Unit type and bedroom count. Timeline to purchase. The agent adapts: a cash buyer ready in 60 days gets routed differently from a financing buyer exploring for next year. Every answer is captured to the CRM in real time.

5.3 Book the gallery visit

For qualified buyers, the agent offers two or three real gallery slots and books one. The appointment lands in the calendar with the qualifying data attached, so the human who runs the visit walks in already knowing the budget, the target unit, and the timeline.

5.4 Tag and hand off

Hot leads get flagged for same-day or next-morning human follow-up. Lower-intent registrations get tagged for the nurture sequence. Nothing falls through. By Monday, the sales team opens a clean, scored, booked pipeline instead of a voicemail box and 260 missed numbers. This routing is standard real estate automation work once the voice layer feeds it.

6. The voice layer vs the alternatives

Developers reach for three other answers to the launch spike. Here is how they hold up against the volume.

Temp staff cannot be hired and trained for a single weekend, and they still answer one call at a time. A form captures a name but holds no conversation, so it never books the visit while intent is hot. A generic overflow center queues callers, which reintroduces the hold time you were trying to kill. The voice layer is the only option that answers everyone, instantly, after hours, without scaling cost per call. That is not a preference. It is the only shape that matches a spike.

7. The window is 6-12 months, not one weekend

Launch weekend is the peak, but a presale runs across construction milestones, not launch-and-leave. So the voice layer is not a one-weekend rental. It absorbs the launch spike, then settles into steady-state inbound for the rest of the sell-out window. The volume of mortgage and contract activity behind those milestones is tracked publicly by outlets like HousingWire, and it does not move in a straight line.

Each milestone fires its own smaller pulse. Topping-out gets a press push and a fresh round of registrations. Facade completion gives the sales team proof-of-progress and a reason to re-engage the waitlist. Every one of those pulses sends inbound after hours, and every one of them deserves a 90-second answer, not a voicemail. The layer you build for launch keeps earning across all of them.

This is why the voice layer pays back over months, not one weekend. You build it once for the worst-case spike, and it handles the long steady tail for free. The marginal cost of the 11th month is near zero. The marginal value of catching one more 300,000 dollar unit in month 11 is another 9,000 dollars of commission.

8. Who this is NOT for

Honest gate, because the wrong fit wastes your money and our time.

This is not for slow-drip developments. If your project trickles in under 10 inquiries a week, with no concentrated launch and no after-hours wall, a human team handles your volume fine. You do not have a spike. You have a stream, and a stream does not need a parallel voice layer. Spend the budget on better creative instead.

It is also not for developers who refuse to staff the back end. The voice layer books gallery visits and tags hot leads. If nobody shows up to run the visit or call the hot leads the next morning, the system just moves the leak one step downstream. The voice layer is the front door. You still have to be home.

And it is not a closer. It does not negotiate price, handle deep objections, or build the relationship that sells a seven-figure unit. It captures, qualifies, books, and hands off. If you wanted an AI to close high-ticket buyers unattended, this is the wrong tool and that is the wrong plan. The honest version of this system knows its lane. See how we score fit in the case studies, or check your own setup on the audit quiz.

9. Build, timeline, and cost

The figures here are luup's own build numbers, not client outcomes. A launch-surge voice agent ships in 5 days and runs at roughly 1,800 dollars per month. That covers the 90-second SLA, the qualification logic, the gallery-booking integration, and the hot-lead tagging.

If you want the booked appointments and tagged leads to flow into the rest of your stack automatically, the automation layer is a 14-day build, 4,500 to 10,000 dollars per month depending on scope. That wires the CRM routing, the nurture sequences, and the milestone re-engagement pulses. Connectors like Make handle the plumbing between the voice agent, the CRM, and the calendar. The project micro-site that converts the paid launch traffic is a 7-day build. The launch creative that drives the spike runs through the ad factory at 40+ assets per month.

Sequence it like this for a launch: site first, creative in parallel, voice layer live before the ads turn on, automation wired to catch the overflow. Build the response capacity before you spend the media that creates the demand. Doing it the other way, spending the paid budget and then watching the registrations leak, is the single most common and most expensive launch mistake there is. If you want a read on where your current launch setup leaks, the fastest path is the revenue leak heatmap, then send us the launch dates and we will tell you straight whether the spike is big enough to justify the build.