What is AI Hallucination?
A hallucination is when an AI gives a confident, well-written answer that is simply wrong. Nothing is broken: the model is guessing the next likely word, so when it does not have the facts it fills the gap with something that sounds right.
In a general chatbot that is a known limit. In front of your customers it costs you money and trust.
Three Ways a Chatbot Makes Things Up in an Online Shop
Example 1: The Invented Return Policy
“You can return any item within 60 days for a full refund, no questions asked.”
The real policy was 28-day returns, unworn items only, original packaging. The customer screenshotted the chat and demanded a refund 50 days later, and the shop paid up to avoid a chargeback.
Example 2: The Wrong Price Quote
“The bundle with the carrying case is £49.99.”
The real price was £64.99. The shop had just changed its prices, but the chatbot was working from older data, and four customers quoted the wrong price at checkout.
Example 3: The Stock Availability Error
“Yes, the navy size L is in stock and ready to ship within 2 days.”
That size had been out of stock for three weeks. The customer ordered, waited for delivery, then got an out-of-stock email. A one-star review followed.
The pattern: in every case the chatbot had no way of knowing the right answer. It answered confidently anyway, and the customer treated it as official.
Why General AI Models Do This
A general AI model learns from a huge amount of text off the internet. It picks up how returns, prices and stock usually work, but it does not know your policies, your prices or your stock. Asked a question, it will do one of these:
- Write an answer that sounds right, based on what a “typical” online shop policy looks like.
- Fall back on old information from its training data, which can be months or years out of date.
- Mix up details from different products, shops or old versions of a policy.
The model does not know it is wrong. How confident it sounds tells you nothing about whether it is right.
What an AI That Uses Only Your Content Does Differently
This kind of AI works only from the policies and FAQs you upload and approve. Every reply comes from your own documents — returns policy, FAQs, product descriptions, shipping terms — and nothing else.
Ask about returns and it reads your returns policy and answers from it word for word. Ask about stock and it checks your live Shopify stock. Ask something you never uploaded and it says so, then hands over to a person instead of guessing.
The rule: it never answers from general knowledge, only from the content you approved. If it does not have the answer, it says so.
What to Look for When Choosing an AI Support Tool
When you compare AI chatbots for your Shopify store, ask these questions:
- Where do its answers come from? If the answer is “a general AI model” with nothing tying it to your own content, move on.
- Can it read live Shopify orders and stock? Documents you upload go out of date, so it also needs the live numbers from your shop.
- What happens when it does not know? It should hand over to a person. A confident guess is not acceptable.
- Can you review and approve what it will say? You should see exactly what the AI will tell customers before it goes live.
- Is there a record of every change? You should be able to read back conversations, and roll your content back to an earlier version if something is wrong.
How AstroX Solves This
You upload your policies and FAQs, and AstroX answers from them and nothing else. It also reads your live Shopify data, so order status and stock come from the real record. When a question is not covered by your content, it hands over to your team with the full conversation attached.
The result: no made-up policies, no invented prices, no wrong stock answers. Just accurate replies, in your own words, from content you control.
See it answer from your own policies
Book a demo and watch AstroX answer using your real store policies — never a guess.
Book a free demo →