Existing customers · AI follow-up

They didn't say no.
Nobody asked why.

TopSales works out what's actually in the way — hassle, price, or a machine that ends up unused — writes the answers into a mind card, and lets that card change the move: send a link, make a call, or hold off this round. Then it runs the follow-up itself. We're paid only on the increment.

Customer 0371 · live walkthrough Illustrative, not real customer data
Conversation
Mind card
This round's move

2–5% of customers held out at random. We're paid only on incremental net GMV. Conversations and orders on this page are illustrative.

Evidence

Every line on the card sits next to the customer's own words.

This customer said two things: “Let me think. The last one had to be taken apart and washed every day. We stopped using it. Message me after 8pm.” Asking the price doesn't make it a price objection; daily washing doesn't make it a no. So the card sorts three ways — confirmed, inferred, still unknown.

What the customer said

How much is this one?

Let me think. The last one had to be taken apart and washed every day. We stopped using it. Message me after 8pm.

Purchase history

1 machine, 2 years ago · no repeat order since

That's all. Two messages, plus their past orders.

The card built from them
FieldCurrent readingEvidence
Cares aboutWhether daily use is a hassleinferred"taken apart and washed every day"
Afraid ofBuying something that sits unusedinferred"we stopped using it"
Stuck onWhether this model is just as hard to cleanunconfirmedThey described the last one, not this one
Current moodCautious, stallinginferred"let me think"
ShiftFrom asking price to explaining why the last one failedOrder of the two messages
Still unknownWhich cleaning step bothers them most; whether price is also an obstacleNot asked yet

The next thing to confirm is the last row: which cleaning step they least want to do.

Decision

The upkeep worry isn't settled. No order push this round.

“Washed every day” decides what to ask first; “after 8pm” decides when to make contact. Each of the three calls carries its reason — and any new reply can overturn one.

This round's decision
CallDecisionWhy
Push or notNo order push — settle the upkeep worry firstUpkeep is the live issue, but they haven't said which step bothers them most
WhenContact today, 20:10The customer said "after 8pm"
WhoTopSales AI this round — unless the customer names their old repAbout usage habits and upkeep, not price negotiation
Scheduled

20:10 — the AI sends one message: which cleaning step bothers you most?

If the customer says: “Let me talk to the salesperson I had before.”
It goes to that rep, together with the customer's own words, the worries already confirmed, and everything already sent — so nobody starts from scratch.

None of this is final: a new reply from the customer can overturn any of it.

Execution

The customer said “just call me.” The AI called.

A message at 20:10. The customer answers “mainly the milk tube,” then adds “I only drink black coffee.” The AI follows both, until the customer asks for the link himself.

20:10TopSales AIMessage sent

"You said the last one had to be taken apart and washed every day. Was the worst part the milk tube, or the grounds and drip tray?"

20:12Customer

"Mainly the milk tube. I only drink black coffee. Just call me."

20:13TopSales AICall placed · 2m 41s

He drinks black coffee and it's the milk tube, so that's what the AI covers:

AI"You drink black coffee, so here's what black coffee leaves you: empty the grounds, rinse the drip tray. Use the milk line and you clean that too. Is that acceptable day to day?"

Customer"Yes. I just don't want to take the milk tube apart every day. Send me the link."

20:16TopSales AIMessage sent

"Here's the model we just went through, with the cleaning notes so you can check them."

Dialogue is illustrative; product facts follow the brand's confirmed cleaning notes for this model. Four things went right: ask which step hurts → say plainly what it takes → wait for acceptance → let them ask for the link.

Update

Once this one is paid, the AI stops chasing.

“Mainly the milk tube” names the worry; “I only drink black coffee” fills in the habit. Those words go back on the card. The order reads paid, so the next step becomes: stop chasing this sale.

Same card · before / after
FieldBeforeAfter
WorryUpkeep — which step, unknownMilk tube is the dealbreaker — "mainly the milk tube"
HabitNot yet knownBlack coffee only — "I only drink black coffee"
AcceptableNot yet knownAccepts the cleanup just described — "yes"
Intent"Let me think" — still hesitatinginferredAsked for the link — "send me the link"
20:28 · Order status

Paid

Card updated to "purchased"; refunds and cancellations stay tracked.

Next step: stop chasing this sale.

Two other endings, written to the card the same way

Customer just says no: record "declined, reason unconfirmed" and stop pushing this model. Upkeep only goes on the card if they say so.

Customer goes quiet: record "no reply, worry still unconfirmed" and re-judge whether and when to follow up — silence is not the same as no interest.

Paid orders stay tracked for refunds and cancellations. What counts is the net GMV left behind; replies only steer the next step.

Increment & pricing

No positive increment, no cut.

From an agreed customer list, 2–5% are held out at random and keep running as before; the rest are followed up with TopSales. Same period, same refund and cancellation basis, then compare net GMV per customer between the two groups.

Holdout · 2–5%

Business as usual

No TopSales. Net GMV per customer recorded as-is.

TopSales · everyone else

Followed up by the model

AI decides whether to push, when, and who. Net GMV per customer recorded.

Incremental net GMV =(net GMV per customer, TopSales − net GMV per customer, holdout) × customers in TopSales group

We take an agreed share of positive increment only. No increment, no cut.

The sale above explains the follow-up; it is not proof of increment. Increment is estimated from the randomised holdout — with a small base or wide order values, results will swing. Observation window, refund basis, our share, and how swings get settled all go into the contract up front.

Start

Take one group of existing customers. Test whether it sells more.

We agree the eligible customers, the holdout, and the settlement terms before anything starts. No system access needed — an export of past conversations is enough.

Built by renlab · hello@renlab.ai

Work phone or email — either one is enough. You'll hear back within a week.

By submitting you agree to be contacted about this enquiry. Your details are used for this conversation only.