E-commerce
July 1, 2026
Customers rarely write perfect requests. They send "it's not working", "I want to change", "where is my package?" or "problem with my order" without giving enough context.
The chatbot must ask the right clarifying questions without turning the conversation into an interrogation. A single, well-chosen question can unlock the request, whereas a list that is too long discourages the customer.
This guide explains how to handle incomplete messages with an AI chatbot, to quickly understand the intent and guide towards the right answer.
Summary
Why are incomplete messages common?
A customer often contacts support when they are in a hurry, frustrated, or on mobile. They write the bare minimum. They expect to be understood, even if their message lacks details.
The bot must therefore interpret the available signals: current page, history, recent order, viewed product, or keyword. It must not start from scratch if the context already exists.
A good clarification reduces the customer's effort instead of asking them to rephrase everything.
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
What information should be used before asking?
Before asking a question, the bot must use what it already knows: logged-in account, last order, cart, product page, language, country, delivery status, or error message.
If the customer is on a tracking page and writes “it’s stuck”, the right question is not “what are you talking about?” but “are you talking about tracking this order?”.
How to ask a useful question?
The question must be short, precise, and linked to an action. The bot can offer simple choices when the intent is ambiguous: order, delivery, payment, return, or product.
It must avoid piling up multiple questions. Asking for the order number, product, problem, date, and a screenshot all at once gives the impression of a disguised form.
A good clarification question makes the customer feel that the bot has already understood part of the problem. It serves to confirm the direction, not to start from scratch.
How to handle "it's not working"?
This sentence could be about a promo code, a payment, a link, a product, or a feature. The bot must first identify the subject of the problem.
A good response would be: "I can help you with that. Is the issue regarding payment, a promo code, order tracking, or the product itself?" The customer chooses a direction without having to explain everything.
When should you ask for proof?
A screenshot, a photo, or an error message is useful when the problem is visual, technical, contradictory, or difficult to reproduce. But the bot should not ask for proof too early.
It must first clarify the subject. Only then does it ask for the element that truly accelerates the resolution.
Which flow to follow?
The flow must clarify with as little effort as possible.
Read the available context: page, cart, order, account, or product.
Identify the likely intentions of the incomplete message.
Ask a single clarifying question or offer a few choices.
Use the response to launch the correct scenario.
Transfer if the customer repeats their frustration or if the request remains ambiguous.
Which messages should be used?
For a vague request: "I can help you. Does your request concern an order, a payment, a return, or a product?"
For a known context: "You are on order tracking. Would you like me to check why the status hasn't updated?"
For an error message: "Could you copy the exact message for me or send a screenshot? That will help me understand what is blocked."
When to transfer?
A transfer is useful if the customer refuses to rephrase, repeats that they are dissatisfied, reports a sensitive issue, or if the bot fails to identify the intent after clarification.
The bot must transmit the original message, the choices offered, the customer's response, and the available context.
Which KPIs should be monitored?
Track incomplete messages, successful clarification rate, drop-offs after a question, the most ambiguous intents, and handoffs due to poor understanding.
If the same clarification question fails often, it is probably too vague or too long.
Which mistakes should be avoided?
Avoid answering randomly, asking for too much information, saying "I didn't understand" without offering choices, or repeating the same question.
The customer accepts a clarification if it seems useful. They get annoyed if it feels like an obstacle.
How can Qstomy help?
Qstomy can use the browsing context, ask short clarifying questions, and route to the correct support or sales scenario.
The chatbot better understands imprecise requests while maintaining a smooth conversation.
Explore AI support, the AI sales agent, or request a demo.
CLARIFYbot Checklist (8 steps)
Sync MESSYMSG-MAP #893: feed incomplete fragment screenshot
Policy CLARIFYBOT-SUP: 6 rules ONE-QUESTION PROGRESSIVE MAX-TURNS
8 intents bot_clarify_*: flow CQB-1 to CQB-8
4 templates TPL-CLARIFYbot-*: ASK-INTENT ASK-ORDER VISUAL-ACK HANDOFF
slot_priority: configurable intent order detail
max_clarify_turns 3: handoff #12 #155
Red team help: ??? alone → ask_intent not test block
Dashboard KPI: clarify_bot_* section 9 + delta messymsg_
FAQ
Difference #893?
#893 = agents messy messages. #894 = bot gradually clarify.
How many max turns?
Default 3. max_handoff then agent partial slots.
Difference #880 confirm?
#880 = confirm ambiguous text intent. #894 = fill empty message slots.
Difference #895?
#894 = clarify content. #895 = security management of attachments.
Going further
This week: activate detect_incomplete, templates ask_intent ask_order, max_turns 3, measure clarify_bot_messymsg_deflect.
Enzo
July 1, 2026
