- Leads decay in minutes, not days — the first five minutes decide who gets the conversation.
- Call first, text and email in the same minute if no answer, retry inside legal windows, then stop.
- Coverage is a software problem. Judgment is yours. Split the job that way.
You paid for the lead. Zillow, your website, a sign call — doesn't matter. The moment it lands, a clock starts, and it's not your clock. It's the lead's. They're on their phone, they're motivated right now, and they filled out three other forms while they were at it.
The industry has known the shape of this curve for years: contact a lead in the first five minutes and your odds of a conversation are dramatically higher than at thirty minutes. Wait a day and you're not following up anymore — you're cold calling someone who already toured the house with somebody else.
01The uncomfortable audit
Before you fix anything, measure it. Export last month's leads and answer three questions for each:
- 01 How many minutes until the first call attempt? (Not text. Call.)
- 02 How many attempts before you gave up?
- 03 What happened to the leads that arrived after 6 PM?
Most teams that run this audit find the same three things: the median first call happens hours later, most leads get one attempt, and evening leads effectively get zero. That last one hurts most, because evenings are when buyers browse.
- lead arrives ─────────── 0:00
- average team calls ───── 4:07:00
- ava calls ────────────── 0:08 ✓
Illustrative timing from our own product behaviour — your market's averages will vary, and anyone quoting a universal number is selling something.
02The play that works
Whether a human or an AI runs it, the winning pattern is boring and repeatable:
Call first, within minutes. A text is easy to ignore; a ringing phone isn't. The first call should come while the lead is still on a listings site.
Fall back in the same minute. No answer? Text and email immediately — reference the exact property, offer two showing slots, not "any time that works!"
Retry on a schedule, then stop. Two or three attempts across the next day, inside legal calling windows, then park the lead in a nurture lane. Persistence wins deals; harassment loses licenses.
Speed-to-lead isn't a talent problem. It's a coverage problem — and coverage is exactly what software is for.
03Where the AI fits (and where it doesn't)
An AI phone agent is very good at exactly the boring parts: being awake at 9:40 PM, dialing eight seconds after the form fill, asking the same four qualifying questions politely for the four-hundredth time, and writing the notes down. That's the coverage layer.
What it shouldn't do is pretend to be you, wing answers about commission, or keep dialing past the attempt cap because a number "looks promising". In our product those aren't guidelines — the compliance gate runs in code before every dial, and Ava discloses she's an AI when asked. We think that's table stakes for anyone doing this seriously.
The handoff is the whole game: the AI covers the first five minutes, and you walk into a conversation that's already qualified, already scheduled, already documented. Your talent goes where it matters — the showing, the negotiation, the relationship.
04Do this today
Run the three-question audit on last month's leads. If your median first-call time is over five minutes — and for most teams it is — fix coverage before you buy another lead. More leads into a leaky funnel is just a bigger puddle.