Employee · Nika
AI win-back agent for online stores
Customers who bought once and vanished — Andrew finds them, messages them with a warm reason to return, and books the repeat order.
A customer bought a coffee machine filter six months ago and never came back. They are not unhappy — they simply forgot. Meanwhile a competitor’s retargeting ad shows up in their feed every week. Without a nudge, that customer is gone for good; with the right message at the right time, they reorder.
Andrew scans your order history for customers whose last purchase is six months or more in the past. He sends each one a short, personal message — referencing what they bought and offering a reason to return. When they reply, he confirms the order details and books it. You see a summary of who came back and who said no, and you pay only for reactivated customers.
How it works in your business
- Finds every customer whose last order is six months or older and messages them personally
- References what they bought and offers a warm reason to reorder — not a generic discount blast
- Books the repeat order when they reply and notes who declined and why
What a dormant customer list is actually made of
Export your orders and sort by last purchase date, and the “inactive” half of the list turns out to be several different groups wearing one label. There are consumable buyers — filters, food, cosmetics — whose product ran out months ago. There are gift buyers who were never shopping for themselves. There are seasonal customers who appear every December. And there are one-time buyers who had a fine experience and simply never thought of you again.
A single “we miss you” campaign treats all of them identically, which is why such campaigns underperform: the message that suits a lapsed consumable buyer is irrelevant to a gift buyer. The information needed to tell them apart is already in your order history — what they bought, when, how often — it just never gets used.
Andrew works from exactly that data. The consumable buyer gets a restock nudge timed to the product’s life; the seasonal buyer hears from him before their season, not in the dead middle of the year; the one-time buyer gets a message that references the actual product they own.
The first month and what it costs
Week one is scoping: you connect the order history, agree the exclusions — recent complaints, unsubscribes, anyone who ordered recently — and approve the message tone with a handful of drafts before anything is sent. Most stores start with their oldest safe segment, where there is nothing to lose.
Weeks two to four are the first real pass through the list. Andrew works through the segments gradually rather than blasting everyone on day one, so replies arrive at a rate you can actually handle. His end-of-week note shows who came back, who declined, and — often the most valuable part — why they declined, in their own words.
The cost only exists where the result does. At price announced at launch per reactivated customer, a pass through a thousand dormant contacts that wakes up a modest fraction of them costs exactly that fraction — the many messages that led nowhere cost nothing. Set that against the margin on a repeat order and the arithmetic settles itself.
What Andrew costs
Pricing is announced when Andrew goes live. Reserve a place and we'll tell you first.
Questions about Andrew
How does Andrew decide who to message?
He looks at the last-order date in your store. Anyone whose most recent purchase is six months or older is a candidate. He excludes customers who unsubscribed, requested no contact, or placed a complaint in the last 90 days.
What does a reactivated customer cost?
price announced at launch per reactivated customer — meaning someone who actually placed a repeat order after Andrew’s message. Messages that got no reply, or replies that did not convert, are not charged.
Does he send discount codes to everyone?
No. Andrew references what the customer bought and offers a reason to return — a restock reminder, a compatible accessory, a new variant. A discount code is only included if you tell him to use one for that segment.
Which channel does Andrew use to reach customers?
Whichever channel the customer left with a valid consent — usually email, with messaging apps where the store already uses them. He respects channel preferences per customer and never hunts for alternative contact routes when one channel does not answer.
We already send newsletters. Does this collide?
No — it is a different motion. Newsletters broadcast to everyone at once; Andrew writes to individuals about the specific thing they bought. You can keep both running: he simply skips anyone your newsletter has recently converted, so no customer is worked twice.
What if he messages someone who just bought, or who had a bad experience?
Recent buyers are excluded by the last-order rule, and complaint flags exclude unhappy customers from the start. If a bad case slips through anyway, the reply lands in his log the same day and you see it — nothing simmers unseen in an inbox nobody checks.