E-commerce
June 30, 2026
Difficult cases rarely occur in high volume, but they have a strong impact: double billing, disputed packages, hacked accounts, highly dissatisfied customers, data deletion requests, or dangerous products. Handling them at random creates costly mistakes.
A library of difficult cases helps agents and the chatbot recognize sensitive situations, collect the right evidence, and escalate at the right time.
This guide shows how to document these cases without turning support into rigid responses.
Summary
Why document difficult cases?
A difficult case often mixes urgency, emotion, money, security, or personal data. If the agent has to improvise, the customer may receive a slow, contradictory, or incomplete response.
The library provides a framework: what to check, what to say, what not to promise, and who should decide.
A difficult case deserves a calm, documented, and consistent response.
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 cases should be included?
Include disputed payments, chargebacks, fraud, hacked account, package delivered but not received, dangerous product, complex warranty, privacy request, aggressive customer, contradictory promise, and out-of-policy commercial gesture.
You must choose the cases that create the most risk, not just the most frequent ones.
How should each case be structured?
Each sheet must explain the input signal, the questions to ask, the necessary evidence, the initial response, the limits, the mistakes to avoid, and the escalation path.
It must also indicate the data that should never be requested in the chat, such as passwords, bank codes, or full card numbers.
How can you help the customer without overpromising?
The client must feel that the matter is being taken seriously, but no promise of refund, compensation, or decision should be made before verification.
The initial response must acknowledge the situation, explain the next step, and prepare for the transfer if necessary.
How do I maintain the library?
Every major incident must enrich the library: new fraud, new carrier-related issue, misunderstood rule, or missing evidence. The sheets must be reviewed with the teams involved.
A static library eventually becomes a source of error.
The sheet must also specify the holding statement to be used when a decision takes time. A client is more accepting of a delay if you explain what is being verified and why an immediate response would be risky.
This transparency limits follow-ups and protects the agent against improvised promises.
Which flow to follow?
The flow must transform the risk into a clear procedure.
Identify high-impact cases: payment, security, privacy, litigation, emotion, or danger.
Describe signals, evidence, questions, limits, messages, and escalations for each case.
Define what the chatbot can collect and what must remain human.
Train agents to quickly recognize signals that fall outside the standard framework.
Update the library after incidents, errors, new rules, or agent feedback.
Which examples should be used?
A "double debit" sheet must distinguish between bank authorization, duplicate orders, and actual debits. A "hacked account" sheet must prioritize security, access, unknown orders, and transfers.
A "parcel delivered but not received" sheet must request proof of delivery, simple verifications, and a carrier inquiry.
When to transfer?
Transfer is necessary as soon as a case involves money, security, personal data, a disputed promise, a dangerous product, strong emotion, or an out-of-bounds decision.
The bot must transmit the signal, evidence, history, risk, urgency, the rule consulted, and customer expectations.
Which KPIs should be monitored?
Track difficult cases handled, errors avoided, escalation times, reopenings, satisfaction, amounts at stake, security incidents, and updated sheets.
This data shows whether the library is actually reducing risk.
Which mistakes should be avoided?
Avoid documenting only the easy cases, hiding limitations, leaving templates without owners, or turning the library into cold scripts.
The framework should support empathy, not replace it.
How can Qstomy help?
Qstomy can connect the chatbot to difficult cases, escalation matrixes, response templates, security rules, SLAs, orders, payments, and support procedures to answer clearly, then transfer sensitive cases with an actionable summary.
The chatbot helps the customer move forward without inventing a sensitive decision, a guaranteed response time, a refund validation, a proof of safety, or an escalation that still needs to be confirmed by a reliable source.
Explore AI support, the AI sales agent, or request a demo.
Key takeaways
Takeaways
A library of difficult cases must cover signals, evidence, limits, initial responses, escalations, and errors to avoid.
What the client must understand
The client must receive a consistent response even in sensitive situations.
The chatbot's correct limit
The chatbot can recognize and collect, but it must transfer money, security, privacy, disputes, and strong emotions.
Enzo
June 30, 2026
