A considered purchase, decided by unanswered questions
A steam cleaner is not an impulse buy. It costs a few hundred euros, replaces a cupboard of chemicals, and has to work on your floors, your grout, your sofa. So before checkout, Techweise shoppers arrive with specific, practical questions, and they ask them in Dutch, French and English:
- Which steam cleaner is right for me?
- What makes the TS2 more powerful than the TS1?
- Can I use it on wood, laminate or vinyl?
- Is steam safe for grout, sealants and the bathroom?
- Can I clean a mattress or sofa with it?
- How do I set it up, and how do I maintain it?
A product page can’t answer all of that for every home. Every question left open is a shopper who keeps researching somewhere else, and often buys somewhere else.
An assistant trained on the catalogue, and on the manuals
So Techweise put Twini on every product page: an AI assistant that answers in the moment, in the shopper’s language. It compares models honestly, says which surfaces each cleaner handles and how, and guides the doubt that matters most in this category, “which one is right for my home?”, like the best salesperson in an appliance store.

The difference is what it’s trained on. Beyond the catalogue, Twini reads the actual TS1, TS2, TS3 and TS Compact manuals, so setup, maintenance and safety answers come straight from the source, not from a generic script.

And the conversation starts early: a chatbar sits on the home page, so shoppers can ask “which one is right for me?” before they’ve even picked a product.

Thousands of conversations, straight to the cart
In three months, Twini handled 6,842 conversations and answered 15,853 times, recommending products 7,826 times across Techweise’s ten markets. The intent shows up in the numbers: 1,346 products added to cart from chat, about one conversation in five, and 847 products sold directly through Twini. On a considered purchase, answered questions convert.

“Shoppers arrive with ‘which one?’ and leave with an answer. Twini gives every visitor the conversation we’d want to have with them in a store, and the numbers followed.”
Tymen WanderCustomer Service Lead at Techweise
Useful after delivery, and a content plan from the chats
Two things make this setup compound. First, the training. Because Twini reads the product manuals, it stays useful after the box arrives: how to descale, which cloth for which floor, what’s safe on natural stone. The same assistant that closes the sale answers the owner’s questions, which means fewer support tickets and fewer avoidable returns.
Second, the listening. Every conversation feeds Insights, which groups what customers keep asking into recurring questions, ranked by volume. For Techweise it surfaced 57 recurring questions, led by “What surfaces and materials can I clean with this steam cleaner, and how?”, asked in 108 conversations, with safety, model comparison and maintenance right behind. It reads like a content plan written by the customers: the guides to publish, the specs to clarify, the ad angles to test.

Together they close the loop: the manuals make Twini a specialist, the conversations tell Techweise what shoppers still need, and the store gets better every week it runs.


