Seven shop questions worth giving to your chatbot
Use seven questions from a cookware shop to decide what AI can answer, what needs clarification and when to bring in a colleague.

Someone comparing two pans rarely wants to “engage with a chatbot”. They want to know whether the handle fits in their oven. The best first content for a shop assistant is often ordinary information your buyers already trust: measurements, care instructions and a delivery policy.
A small shop’s question sheet
The examples below describe a fictional cookware shop. Give your team the same task with a product they know well. For each question, write down the public source and the point at which a person should decide.

| Visitor’s question | Best next step |
|---|---|
| Does the 24 cm pan work on induction? | Read that model’s specification and cite it. |
| Can the lid go in the dishwasher? | Use the lid’s care page. The pan may have different instructions. |
| Can I collect on Saturday? | Give the published collection hours and the correct location. |
| Where do I send an unused return? | Link the applicable returns instructions so the visitor can read the conditions. |
| Will it fit in my oven? | Ask about usable depth, then compare it with the total length including the handle. |
| Where is my individual parcel? | Route to the order-support team through the shop’s usual process. |
| Can you replace this damaged pan today? | A person checks the damage and available remedies. |
Why a small detail in the source matters
A diameter of 24 cm describes the cooking surface. It says little about the overall length. If the total length is missing, a confident “yes” can leave a buyer with a pan they cannot use. Add the dimension to the published specification, refresh the knowledge source and repeat the original question. The next visitor who asks in different words benefits from the same correction.
Pick one category for your first week
Enter the category’s public pages in QuatroChat, install the widget once and test the four factual questions. Read the cited page on a phone. Then ask the two exception questions in Hybrid mode. In the shared inbox check that the colleague sees the exact model and the history before replying. If the team prefers to handle every conversation while refining the sources, it can use free live chat throughout.