Case study
A buyer asks an AI assistant which power bank can charge a laptop and a phone at the same time on a business trip. The product fits. Its product information doesn’t show it. This is what that gap looks like, and what it takes to close it, from the answer to the order.
Anonymized example built from a real buyer question. Brand, model and figures are illustrative, not a customer result.
“For business travel, which power bank can charge my laptop and my phone at the same time?”
It reads like one question. It is three conditions, and a product has to meet all three before an assistant can name it with confidence:
The product was a reasonable fit on all three. Its page and its marketplace listing didn’t make that checkable. Illustrative figures:
| The buyer needs to know | What was published | Status |
|---|---|---|
| Can it go in carry-on? | “20,000mAh high capacity.” No watt-hour rating anywhere. | Missing |
| What does the laptop get with a phone plugged in? | The page said “100W laptop charging”; the marketplace listing said “65W + 65W.” Both were true, in different modes. Neither said which applies when two devices are connected. | Conflicting |
| Is it travel-safe? | “TSA approved.” TSA does not certify products, so the claim has nothing behind it. | Unconfirmed |
| Is it the wrong choice for my laptop? | Not stated. | Missing |
A headline figure of “130W total output” sat above all of this. It was accurate, and it answered none of the three conditions.
An assistant answering this question can only go on what it can read. It may apply the travel rules more strictly than the regulations do, for example by setting aside any power bank whose rating it cannot confirm, or read “130W” as more than the laptop will actually get. Its reading may be right or wrong. Either way, the product is not in the answer, and the brand cannot argue with that reasoning after the fact.
No assistant publishes how it makes that call. What the brand controls is the product information: state the facts the question depends on, in the units the question is asked in, so nothing is left to guess.
After the change, a buyer who asks this question can get the same answer wherever they check: in the AI assistant, on the marketplace listing they search next, on the product page, and from a sales rep. That matters because the assistant is often where the question starts, not where the purchase happens.
The follow-up is to ask the same question again in AI answer checks and compare the answers, and to watch AI referral visits in Google Analytics 4. Feeding inquiries, deals and orders back into the next round of priorities is part of lead-to-contract feedback (coming in v3.0).
What this case does not claim: that any assistant will now recommend the product. That is still the assistant’s decision. The product is no longer ruled out by facts its own page left out.
What the case shows. A buying question is rarely about one keyword. It carries conditions the buyer may never say out loud: that the product can fly, that it can power the laptop, that it can do both at once. Hasmord works on the gap between those conditions and a brand’s product information, and follows each fix from the AI answer toward the order.
Related: how to find missed buying opportunities · product claims and evidence checks · AEO for Amazon sellers.
Travel rules: FAA PackSafe, lithium batteries; TSA, lithium batteries with more than 100 watt hours. Airlines may set stricter limits.
Hasmord is the merchant operating system connecting AEO to sales. It finds the buying questions your products can’t yet answer across sales and AI shopping channels, turns them into channel-specific improvements your team reviews, and checks what changes afterward. Lead-to-contract sales management and sales feedback are coming in v3.0.