Travelpro Europe tailors its chat prompt to every type of page

A common mistake with on-site AI chat widgets is treating them as a single feature with a single default state: the same greeting, the same suggested prompts and the same behavior, whether the shopper has just arrived on the homepage or is deep into a product's spec sheet. That's easier to build, but it misses an opportunity, because a shopper on a landing page and a shopper on a product page are in very different frames of mind.

Travelpro Europe, using Manifest AI, tunes its nudges by page type instead. On general search and landing pages, the nudge is broad, offering ways to browse by luggage style, a seasonal discount prompt and a question about site benefits, since the shopper hasn't narrowed down what they want yet. On collection pages it's lighter, with a general "ask me anything" and a bestsellers shortcut, because the shopper has already chosen a category and doesn't need the full menu again.

Product pages get the most specific treatment. On a spinner suitcase page, the nudge offers three questions generated from that product's attributes, the things a shopper looking at that particular item would plausibly want to know. On a backpack page without pre-built attribute questions, it falls back to a general "know more about this product" prompt with a similar-products option, so the shopper always has somewhere to start.

The principle behind all four variants is the same: keep shoppers on the page they're engaged with, instead of sending them elsewhere for answers. A shopper who has to leave a product page to find information may not come back to it, and a nudge that answers the question in place removes that risk.

In the last 30 days, 72,863 shoppers were nudged on product pages specifically, converting at 0.68%. It's important to read that number in the right context instead of comparing it with a benchmark that doesn't apply. This isn't a checkout nudge or a cart recovery message aimed at someone who has already decided to buy. It's an engagement layer earlier in the funnel, reaching shoppers who are still browsing and haven't necessarily formed purchase intent. A 0.68% conversion rate from a broad, browse-stage touchpoint is a different kind of number from a 20%+ rate on a bottom-of-funnel cart nudge, and comparing them directly would be misleading.

What the volume does show clearly is reach. This runs at real scale, across tens of thousands of product-page visits a month, without anyone needing to decide what each page should say. The tuning by page type happens once, during setup, and then runs on its own across every relevant page.

If your on-site chat widget shows one default prompt everywhere, this is the gap to close. Check whether the nudge on your product page knows which product it's sitting on.

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.