# Customer Questions and Support

*AI for E-commerce*

Where is my order, does it fit, can I return it

The overwhelming majority of store support is three questions repeated forever: where is my order, will this fit or work with what I have, and how do I return it.,Your employee answers them from your real data. Order status from the store, sizing and compatibility from the product information, returns from your actual policy rather than a generic one.,What is left after that is the genuinely unusual, which is the part that deserved your attention in the first place.

## Benefits

### undefined



### undefined



### undefined



### undefined



## How It Works

1. **Step 1**: 
2. **Step 2**: 
3. **Step 3**: 
4. **Step 4**: 

## At a Glance

- **Live** Order data behind every answer
- **3** Questions that are most of the volume
- **Routed** The unusual reaches a human
- **Fed back** Repeat questions fix the listing

## Three Questions Are Most of the Volume

Store support feels varied when you are in it and is remarkably repetitive when measured. Where is my order, will this fit or work with what I already own, and how do I send it back account for the large majority of contacts in most catalogs. All three are answerable from data the store already holds, which is what makes them worth handling systematically. The reason they consume so much of an owner day is not difficulty; it is that each one requires looking something up, and doing that a hundred times a week is exactly the work a person should not be doing.

## A Repeated Question Is a Listing Bug

If twenty customers ask whether a product fits a particular model, the product page failed twenty times before anyone contacted you, and for every person who asked there were others who simply left. Treating recurring support questions as a queue to answer faster misses the actual opportunity, which is to fix the page so the question is not necessary. That closes the loop between support and the catalog, and it is the difference between support that scales linearly with orders and support that gets quieter as the listings get better.

## Keep Money Decisions on a Human

Order status and sizing are safe to automate because a wrong answer is embarrassing and correctable. A refund is neither. It moves real money, it is awkward to reverse, and the situations that trigger one are exactly the situations where a customer is already unhappy and judging how you handle it. The practical rule is that reading and explaining can be automatic, while anything that moves money out stays a decision a person makes, even when the platform would let you do otherwise.

## FAQ

### Will it reply to customers on its own?

That is your choice per channel. Many stores start with drafts for review and move routine order-status replies to automatic once they have watched it for a while. Anything involving a refund, a dispute, or an unhappy customer is worth keeping in front of a human regardless.

### How does it know where an order is?

It reads the connected store and the fulfilment status there, so the answer is about that specific order. Where tracking has not updated, the honest answer is that it has not updated, which is more useful than a confident guess about a parcel nobody can see.

### Can it issue refunds?

Refunds move money, so treat them as you would any spend: keep them behind your approval unless you have deliberately set that app to act without asking. The recommendation is to keep a human on money out.

### What about questions in other languages?

It answers in the language the customer wrote in. For a store selling internationally that is often the single biggest support improvement available, since the alternative is usually a slow reply or none.