The date that triggers it
The last purchase, the last delivery, the last payment — or the custom field you define. It's calculated at send time, not when you built the rule, so whoever enters the condition today receives it today.
There's no cancellation, no complaint, no call. You simply look at one month's revenue and people are missing. Cobratia detects them with the rule you set —exactly fourteen days without buying, or thirty, or sixty— and writes to them on WhatsApp before they get used to buying from someone else.
* Nothing is claimed here. They don't need to owe you money.
Look at what it doesn't say. It doesn't say "DON'T MISS OUT!", it doesn't send a coupon, it doesn't mention a campaign. It says the name, says what they used to buy and says how long since they last showed up. That's what makes them reply.
The "three weeks ago" and the product weren't written by anyone: they come from their record at send time. That's why the message can be sent to two hundred people and each one receives their own.
And when she replies, she isn't sent to a form or told "an advisor will contact you". The assistant is still there and takes her order in the same conversation.
Daily detection
They stopped buying
Marina Ferraro
31 days without ordering · used to buy every 15 · we wrote
Damián Ocampo
44 days without ordering · replied and ordered
Rocío Baigorria
38 days without ordering · awaiting reply
Tomás Iriarte
Skipped · has debt, collections handles it
Alejandra Puig
52 days without ordering · told us why they left
Ignacio Bordón
Skipped · asked not to receive
Verónica Sandoval
29 days without ordering · we wrote
Hernán Vidal
63 days without ordering · bought again
Anyone can send a promo. What changes the result is writing to the right person at the right moment, and that's decided by the condition you build once.
The last purchase, the last delivery, the last payment — or the custom field you define. It's calculated at send time, not when you built the rule, so whoever enters the condition today receives it today.
A single rule at 30 days is a single chance. What works is a sequence: at 7 a gentle reminder, at 14 a question, at 21 an offer, at 28 the last try. Each with its own text.
As you build the rule, "X of your Y contacts match" appears with a sample. And before launching, you simulate the full batch without sending anything.
If you have products loaded, the message can carry what that customer used to buy, with its price. Not a generic promo you can spot a mile away.
That's where it's won. The assistant continues the chat, and if you have orders enabled it takes the order on the spot, without passing it to anyone or sending them to an app.
Both engines respect the same minimum distance between messages. Nobody gets a debt demand and a repurchase offer on the same day.
This is the sequence running in production today at a real customer. Each milestone has its text and its tone, and whoever replies at the first one doesn't receive the ones that follow.
| When | What they're told |
|---|---|
| 7 days | No selling. An "everything all right?" that opens the door without intruding. Half of those who came back, come back here. |
| 14 days | Naming what they used to buy, with the product and the quantity from last time. It's the one that generates the most orders. |
| 21 days | Only now a concrete reason to come back: an offer, something new, a benefit. Earlier it would have been rushed. |
| 28 days | The last one. Short and without insisting, leaving the door open. After this one, the writing stops. |
The days are an example, not a recipe: at a water distributor delivering weekly, fourteen days is already late; at a furniture store, sixty is early. The sequence is built with you according to how often your average customer buys.
Real numbers from the customers using the platform today, added up and without naming any. They update on their own.
What people ask us most before getting started. If something's missing, write to us and we'll answer it.
From your system, if it sends it mapped in the connector, or from the orders the assistant took. It's the right question and it's worth settling at the start: if a contact doesn't have that date loaded, it doesn't enter the rule. That's not a bug, it's what should happen — we'd rather leave someone out than write to someone who bought from you yesterday.
Any other date field you do have can be used: the last delivery, the last payment, the last visit. Or you start accumulating it once the assistant takes orders. During onboarding we look at it with you and choose which data to start from.
No, if the rule is set properly — and you'll see it before activating it, with the exact number of how many it reaches and a sample of who they are. The simulation also tells you, one by one, why each contact left out is skipped.
Yes, with a template approved by Meta. WhatsApp only allows free-form messages if the person wrote to you in the last 24 hours, and precisely the one who stopped buying a while ago is outside that window. You write the templates from the panel and they're submitted for approval on their own.
We won't estimate it for you, and be wary of anyone who throws a percentage at you without knowing your business. It depends on your industry, how long ago they left, why they left and what you offer them. What we do is measure it from day one: the system counts those who bought again after being inactive, not the messages that went out. In two months you'll have your own number, which is worth more than any average we could give you.
It's the same engine, with one difference that matters: a campaign speaks to an audience you choose by whatever criteria, and this speaks to one person because of their history — it's been exactly so many days since they bought from you. That's why the message can be personal instead of a promo for everyone.
That's the awkward situation as always, and it's covered: both engines share the minimum distance between messages, so they don't get the reminder and the offer on the same day. You decide which goes first.
Everything comes from there. With that number we build the sequence and define after how many days someone is "lost" for your business. And we check whether you have the last-purchase date loaded — because if you don't, that gets solved first.
Winning back is one of the five things the platform does.