Your second brain
for closing deals.
Speak after a showing. Forward an email. Pull up a client. Lumi captures the soft signals, fills the brief, and feeds Claude — automatically.
- Voice → CRM, auto. No forms.
- Works offline. Syncs when you're back.
- Free for agents in EU · LatAm · MENA.
9-min read · Updated April 2026
Lumi · Wednesday
Good morning, Niki.
Two showings · three leads need a nudge.
Showing · Passeig de Gràcia 84
The 7-day silent buyer.
One sentence re-opens 40%.
Most agents lose silent buyers to a template. “Just checking in.” “Wanted to follow up.” “Are you still looking?” The buyer reads it in two seconds, files it under “I'll reply later”, and never does. The agents who re-open the most cold buyers are doing something different — and the difference fits in one sentence.
Sofia — saw a place this morning with the same kind of kitchen window you liked at the Rua da Prata showing. Ground floor, balcony. Want me to send the link? — A.
Three sentences. No mention of the silence. One specific detail. One 60-second next step.
Why the templates fail.
A silent buyer is not a buyer who has lost interest. They're a buyer whose interest has been overtaken by something — a work deadline, an argument with a spouse, a different listing they saw on a Saturday and haven't mentioned to you. The window of attention you had after the last meeting has closed. To re-open it, you need to give them a reason to look up.
“Just checking in” is not a reason. It's a request. It puts the burden on the buyer to remember where things stood, summarise their current state, and decide whether they want to spend energy answering. The cost of replying is high. The cost of ignoring is zero. They ignore.
The reactivation message that works inverts the cost equation: it costs nothing to reply, and the message itself rewards them for opening it. It references something they cared about. It surfaces something new in the world. It asks a yes/no question that takes 6 seconds. The whole interaction is a gift, not a request.
“The reactivation message is a gift, not a request. Templates flip the cost equation the wrong way.”
Five rules. One message.
Each rule is a constraint that filters out a common failure mode. Skip any one and the message regresses to the template that doesn't work.
- 01
The trigger fires at day 7, not day 3 or day 14.
Day 3 is too early — buyers are processing. Day 14 is too late — momentum is gone, and a longer silence demands a different message (the 'graceful exit' message, not the 'reactivation' one). The 7-day mark is the honest sweet spot: enough silence to mean something, not so much that you've lost the right to a casual reach-out.
- 02
The brief gets pulled, not the lead.
The whole point: the AI is reading the 7-field brief, not the lead's name. If your CRM doesn't have soft_signals captured, the message regresses to 'just checking in'. The reactivation message is downstream of the brief — that's why the agents who do this well are the same agents who voice-note their soft signals on the walk back to the car.
- 03
Claude drafts. You don't send.
The agent's job is one tap of approval — or one tap of edit. The model gets it right ~70% of the time on first pass. The other 30% needs a 5-second human edit (a name correction, a tone tweak, a different soft signal swapped in). Never auto-send. The whole credibility of the message is that it sounds like you, written for them.
- 04
Same channel as last touch.
If the last touch was a WhatsApp voice note, the reactivation goes over WhatsApp text — not email. Channel-shift signals desperation (or worse, a CRM-driven sequence). Continuity of channel is part of what makes the message land as personal.
- 05
If they don't reply in 48h, you stop.
One reactivation message. That's it. If they don't reply within 48 hours, the next touch is 6 weeks out and is a market update, not a follow-up. The reactivation works because it's rare. Sending a second one in week 2 burns the trick.
The 6-week reactivation clock.
The exact cadence the protocol runs on. Set the trigger once and let it fire — but understand each step, because the moments matter.
What specificity sounds like.
Same buyer profile. Different soft signals captured at the showing. Same prompt produces four completely different reactivation messages.
“Sofia — saw a place this morning with the same kind of kitchen window you liked at Rua da Prata. Ground floor, balcony. Want me to send the link?”
“Quick one — building came up in Lapa with three families on the same floor, kids ages 4-9. Two-bed, ground floor. Worth a look this Saturday morning?”
“Spotted a place 25 min from Estoril, walkable to the train. South-facing balcony. Sending the link if you want it — Saturday viewing slot is free.”
“Two ground-floor options came up in your range this week — both with the balcony you wanted. Sending Thursday like I said. Want me to bump it to today?”
Notice what's consistent: each message is 1-3 sentences, references one specific detail from the brief, anchors in something new in the world (a listing, a building, a viewing slot), and ends with a yes/no question. Notice what's absent: no “just checking in”, no “hope you're well”, no emoji, no exclamation marks, no offer of a 30-minute call.
Three messages the model defaults to without a brief.
When the prompt is given just a name and 'days_silent: 7' without the soft_signals layer, AI regresses to one of these three failure modes. Each one is a template wearing different clothes.
“Hi Sofia! Just wanted to follow up and see if you had any questions about the Rua da Prata apartment. Let me know if you'd like to see anything else!”
Mentions the silence ("just wanted to follow up"). Generic. No reference to anything specific. Reads as template. Buyer files it under 'I'll get to it' — and never does.
“Sofia, I've been thinking about your search and I want to make sure we don't lose momentum — there are 3 great options I think you'll love, do you have 30 mins this week to chat?”
Pressures. Asks for 30 minutes when the buyer just wanted a 6-second message. The 'don't lose momentum' is agent-anxiety leaking onto the page.
“Sofia! Hope you're doing well :) Wanted to share 5 new listings I think might fit your search. Take a look and let me know what you think!”
5 listings is admission that the agent doesn't know which one matters. Hope-you're-doing-well + emoji + 'let me know what you think' is the template stack that AI defaults to without a specific brief.
The brief that powers the message.
This is the input the AI receives. Notice that the soft_signals are verbatim, the next_promised step from the last touch is captured, and the days_silent count is computed automatically.
# ── client brief — input to the prompt ────────────
client: "Sofia Ferreira"
last_touch:
at: 2026-04-19 09:14
channel: "WhatsApp voice note"
content: "follow-up after Rua da
Prata showing — flagged
the second-floor stairs"
next_promised: "send 2 ground-floor +
balcony options by Thu"
soft_signals:
- "paused at the kitchen window —
'this is where I'd make coffee'"
- "asked twice if neighbours had kids"
- "mother lives in Estoril"
intent_stage: "serious"
window_earliest: 2026-07-01
window_latest: 2026-10-15
days_silent: 7
The full 7-field brief structure — what each field captures, why it matters — is documented in our prompt guide. The reactivation message is one of the workflows that field guide unlocks.
What to feed Claude.
The system prompt that turns the brief into the message. Tested against Claude Haiku and Sonnet — Haiku is fast enough for real-time and produces the right voice on first pass.
You are a senior real-estate agent's reactivation drafter.
INPUT
You will receive: client name, last_touch summary,
soft_signals list, intent_stage, window dates,
days_silent count.
OUTPUT
Write ONE message — never two — to be sent over
the same channel as last_touch (WhatsApp, SMS,
or email). Length: 1-3 sentences. No subject line
unless email.
RULES (non-negotiable)
1. Reference exactly ONE specific soft_signal
from the brief. Not paraphrased. The detail
itself ("the kitchen window", "the south-
facing balcony", "your mother in Estoril").
2. Do NOT mention the silence. Never say
"wanted to check in", "haven't heard back",
"just following up", "circling back".
3. Anchor the message in something that has
changed in the world since last_touch:
- a new listing that hits soft_signal
- a price drop on something they saw
- a market data point relevant to window
- a personal detail (birthday, season,
holiday) that fits the soft_signal
4. End with a single concrete next step —
a date, a time, or a question that takes
60 seconds to answer. Not "let me know".
5. Match tone to intent_stage:
browse → casual, no urgency
serious → warm, calendar-aware
urgent → direct, name the window
6. Sign off with the agent's first name only.
No "Best regards". No agency footer.
ANTI-PATTERNS (never produce these)
- "Hope you're doing well"
- "Just touching base"
- "Wanted to make sure my last message wasn't
lost in your inbox"
- "Are you still looking?"
- "Let me know if you have any questions"
- Any emoji
- Any exclamation marks
Voice: warm, brief, specific. The buyer should
read this in 6 seconds and feel seen, not sold.Copy the system prompt above into a new Claude chat as a system message, then paste a brief like the one above as your first user message.
What Claude returns.
Run the brief above through the prompt above. This is the first-pass output — no editing.
Sofia — saw a place this morning with the same kind of kitchen window you liked at the Rua da Prata showing. Ground floor, balcony, 30 min from your mother's place in Estoril. Want me to send the link? — A.
Where the 40% comes from.
The figure is a working benchmark we've seen across 40-odd agents in EU and LatAm using a brief-driven reactivation message. The range is wide — agents with thin briefs hit 18-25%, agents with full briefs and disciplined day-7 timing see 38-52%. The relevant comparison isn't the “industry average” (no such thing for this measure) but the agent's own pre-protocol baseline. Most agents we've worked with were re-opening 8-12% of silent buyers with template messages. Going to 40% is a 3-4× lift on a meaningful denominator (the fraction of leads that go silent at 7 days is typically 25-35% of new contacts).
The honest caveat: “re-opens” means “buyer replies within 48 hours of the message”, not “buyer closes a transaction”. Reply rate is the leading indicator the protocol optimises for. Closing rate downstream is a function of the rest of your pipeline, not this single message.
Drafting the message is step one.
Trusting the trigger is step two.
Lumi is the app that runs this workflow for you. You speak after a showing — Lumi captures the soft signals. You forward an email — Lumi updates the constraints. You open the app at 8am — the brief is already there, ready to feed Claude.
- Voice → structured CRM, automatically
- No forms. No data entry. No copy-paste.
- Free for agents in EU · LatAm · MENA
Lumi · Wednesday
Good morning, Niki.
Two showings · three leads need a nudge.
Showing · Passeig de Gràcia 84
Pipeline
Active
8
Warm
4
Cold
2
Clara Ruiz
Active€1.8M · 3BR
Passeig de Gràcia showing · 11:30
Andreas Moreno
Active€2.4M · 4BR
Send comps by 18:00
Dimitri Schneider
Warm€900K · 2BR
Contract review today
Silent 3d · last 3 days ago
Sarah Mitchell
Cold€1.2M · 3BR
Draft re-engagement
Silent 9d · last 9 days ago
A real-estate adaptation of the reactivation-messaging thesis from vibe-marketing — specificity beats cadence. Our slice: 7-day silent buyers, and the soft signals captured at showings that most agents fail to log.
More guides like this on @lumi.estate. Follow if any of this was useful — it's how we know to keep writing.