What happens when a shopper asks Travelpro Europe a real question about luggage

Product pages are built to answer the questions a merchandising team anticipated: dimensions, materials and a few bullet points on features. What they can't do is answer the specific question a shopper has in mind, the one that doesn't fit neatly into a spec sheet.

Travelpro Europe gets those questions all the time, because luggage decisions come down to details. Can the handle adjust for a shorter traveler? Will the wheels hold up on cobblestones? Is this toiletry organizer actually leakproof, or only water-resistant? Does the backpack's security pocket really block RFID skimming, or is that just marketing language?

Manifest AI answers these live in the chat widget, across the full catalog rather than a single product category. A shopper asking about handle adjustability on a spinner suitcase gets a grounded answer drawn from real product data, instead of a generic deflection to the return policy. The same applies to liquid and gel safety on the organizer and to the RFID-blocking claim on the backpack. Each answer links back to the relevant product page and closes with View Product or Add to Cart, so the conversation has somewhere to go once the question is answered.

The most important number here is resolution: 2,796 of 3,044 queries were resolved automatically in the last 30 days. That's a 92% reduction in the volume that would otherwise reach a human support agent. For a brand selling a considered purchase like luggage, where a wrong answer to a spec question can lose a sale outright, getting that right without a person in the loop is significant.

It's important to be precise about what "resolved" means. It doesn't mean every one of those 2,796 conversations ended in a purchase. It means the shopper received a specific, grounded answer to their question and didn't need to escalate to a person. That's a valuable outcome in its own right. A shopper whose handle question is answered in 10 seconds, rather than after a wait for a support email, is more likely to still be on the page, and still in a buying frame of mind, when the answer arrives.

This can go wrong when an AI chat widget gives confident answers that aren't grounded in a product's real specs, and a shopper makes a decision based on something invented. Travelpro's setup avoids that by training on actual catalog data instead of letting the model improvise. It's the less flashy approach, and the more defensible one. It also means some questions genuinely go unanswered, the ones the training data doesn't cover, and those make up the remaining 8%. Acknowledging that gap is part of what makes the 92% figure credible.

If you're deciding whether an AI chat widget is worth it for a considered-purchase catalog, look past Travelpro's headline resolution rate to the ratio of grounded answers to invented ones, and to how much of your catalog data is structured well enough to train on before you'd trust it live on a product page.

See how this would work on your store

If you want to build a flow like this for your own store, book a meeting with the Manifest AI team and we will walk you through it in a live demo.