E-commerce
June 28, 2026
These messages concern an abandoned cart. The customer is blocked, has doubts, or does not understand what is happening.
The objective is simple: explain the problem in clear words, say what to do next, and avoid vague answers.
Summary
Why use a chatbot to recover abandoned carts instead of email alone?
This problem occurs when an abandoned cart blocks the customer or creates doubt.
Support must first understand the cause, then explain the solution without blaming the customer.
Respond in less than 5 seconds to a specific objection
Tailor the message to the actual content of the cart
Offer a direct payment link without leaving the conversation
Detect a technical block (promo code, stock, non-deliverable zone)
Convert over 2,000 customers on average per month with Qstomy.
The world’s 1st Shopify AI dedicated to customer conversion
Empowering 200+ e-commerce merchants
Which three positive signals trigger the intervention?
The question is not "when to interrupt," but "what objective signal deserves the interruption." ShoplyAI insists: a threshold without a page signal is an approximation that cannot be used in auditing
Signal 1: cart inactivity
Signal 2: intention to leave the page
Signal 3: back-and-forth on policy
Eligible pages
Which two states ban all proactive messaging?
Two negative security rules take precedence over any positive signal. Forgetting them silently destroys conversion
Maximum of one proactive intervention per session on cart/checkout
Respect of "no thank you" and the closed chat window
No signal on converted visitor within 24 hours
Which ready-to-use messages should you use?
Messages must be short, human, and helpful. The customer must know what to do after reading them.
Delivery: "Delivery to [country]: [timeframe] from [amount]. Any questions about delivery times?"
Returns: "Free returns within [X] days. Hesitating about the size or model?"
Payment: "Secure payment, no hidden charges. Any doubts about the payment method?"
Size or compatibility: "The [cart product] is suitable if [criterion]. I can check for your specific use."
Promo code: "Your code seems to be rejected: I'm checking the conditions live."
How do you contextualize with the actual content of the cart?
A generic "how can I help you?" bot without knowing the cart is perceived as a call center
Cart line items titles and variants
Amount and remaining free shipping threshold
Detected delivery country
Logged-in customer: past orders, usual size
Should a discount be offered in the chat?
A systematic discount on the first contact trains your customers to deliberately abandon their carts in order to trigger the coupon
Touch 1 chat: factual response, zero discount
Touch 2 email D+1: cart reminder + return/payment reassurance
Touch 3 D+3 or: discount only if margin and cart justify it
How to orchestrate cascading chatbot, email, and SMS?
The best rates combine cascading channels, with each relay picking up where the last one left off
T+T+5 min: in-session chatbot based on behavioral signal
T+1 h: neutral email if email captured, without discount
T+24 h: reassurance email (reviews, returns, payment security)
T+48-72 h: or incentive email if cart > margin threshold
Which rules should be followed?
Rules serve to avoid vague or false answers.
Do not answer at random.
Verify information before promising a solution.
Explain the situation with simple words.
Clearly state the next step to the customer.
Transfer to a human if the case is risky, legal, or blocked.
How to connect Shopify, knowledge base, and transfer to a human?
The bot must quote your real policies, not improvise
Chat window with configurable signals + negative rules
Help center knowledge base: shipping, returns, payment
Update Shopify: cart, stock, variants, customer
Transfer to a human agent if cart > threshold or frustration detected
Which indicators should be monitored?
You need to track a few simple metrics to know if the support is genuinely helping customers.
Number of requests about this issue.
First-contact resolution rate.
Average response time.
Number of cases escalated to a human.
Customer satisfaction after the response.
How can Qstomy help?
Qstomy can help support recognize the customer's request and find the right information.
The chatbot answers simple cases. If the case is blocked, sensitive, or too complex, it transfers the conversation to a human.
The goal is to respond faster, without inventing answers and without losing context.
What should be put in place this week?
Here is a simple version to implement without complicating support.
Identify the customer's exact problem.
Check the information in the order, cart, subscription, or product sheet.
Respond with simple words and clear action.
Escalate to a human if the case is sensitive, legal, or blocked.
Look at repeated requests to fix the page, payment, or rule involved.
Enzo
June 28, 2026
